noop-length-coercion-v2.patch

text/plain

Filename: noop-length-coercion-v2.patch
Type: text/plain
Part: 0
Message: Re: Identifying no-op length coercions

Patch

Same data as JSON: GET /api/v1/attachments/:id/patch the parsed metadata as JSON — format, series position, per-file stats; never the diff bytes. API reference →
Format: unified
Series: patch v2
File+
doc/src/sgml/catalogs.sgml 0 0
diff --git a/doc/src/sgml/catalogs.sgml b/doc/src/sgml/catalogs.sgml
index 8504555..5d1a447 100644
*** a/doc/src/sgml/catalogs.sgml
--- b/doc/src/sgml/catalogs.sgml
***************
*** 4337,4342 ****
--- 4337,4349 ----
       </row>
  
       <row>
+       <entry><structfield>protransform</structfield></entry>
+       <entry><type>regproc</type></entry>
+       <entry><literal><link linkend="catalog-pg-proc"><structname>pg_proc</structname></link>.oid</literal></entry>
+       <entry>Calls to function can be simplified by this other function</entry>
+      </row>
+ 
+      <row>
        <entry><structfield>proisagg</structfield></entry>
        <entry><type>bool</type></entry>
        <entry></entry>
diff --git a/src/backend/catalog/index 6250b07..92be0a7 100644
*** a/src/backend/catalog/pg_proc.c
--- b/src/backend/catalog/pg_proc.c
***************
*** 304,309 **** ProcedureCreate(const char *procedureName,
--- 304,310 ----
  	values[Anum_pg_proc_procost - 1] = Float4GetDatum(procost);
  	values[Anum_pg_proc_prorows - 1] = Float4GetDatum(prorows);
  	values[Anum_pg_proc_provariadic - 1] = ObjectIdGetDatum(variadicType);
+ 	values[Anum_pg_proc_protransform - 1] = ObjectIdGetDatum(InvalidOid);
  	values[Anum_pg_proc_proisagg - 1] = BoolGetDatum(isAgg);
  	values[Anum_pg_proc_proiswindow - 1] = BoolGetDatum(isWindowFunc);
  	values[Anum_pg_proc_prosecdef - 1] = BoolGetDatum(security_definer);
diff --git a/src/backend/commands/taindex 2c9f855..563a1b2 100644
*** a/src/backend/commands/tablecmds.c
--- b/src/backend/commands/tablecmds.c
***************
*** 56,61 ****
--- 56,62 ----
  #include "nodes/nodeFuncs.h"
  #include "nodes/parsenodes.h"
  #include "optimizer/clauses.h"
+ #include "optimizer/planner.h"
  #include "parser/parse_clause.h"
  #include "parser/parse_coerce.h"
  #include "parser/parse_collate.h"
***************
*** 3495,3501 **** ATRewriteTable(AlteredTableInfo *tab, Oid OIDNewHeap, LOCKMODE lockmode)
  	{
  		NewColumnValue *ex = lfirst(l);
  
! 		ex->exprstate = ExecPrepareExpr((Expr *) ex->expr, estate);
  	}
  
  	notnull_attrs = NIL;
--- 3496,3503 ----
  	{
  		NewColumnValue *ex = lfirst(l);
  
! 		/* expr already planned */
! 		ex->exprstate = ExecInitExpr((Expr *) ex->expr, NULL);
  	}
  
  	notnull_attrs = NIL;
***************
*** 4398,4404 **** ATExecAddColumn(List **wqueue, AlteredTableInfo *tab, Relation rel,
  
  			newval = (NewColumnValue *) palloc0(sizeof(NewColumnValue));
  			newval->attnum = attribute.attnum;
! 			newval->expr = defval;
  
  			tab->newvals = lappend(tab->newvals, newval);
  			tab->rewrite = true;
--- 4400,4406 ----
  
  			newval = (NewColumnValue *) palloc0(sizeof(NewColumnValue));
  			newval->attnum = attribute.attnum;
! 			newval->expr = expression_planner(defval);
  
  			tab->newvals = lappend(tab->newvals, newval);
  			tab->rewrite = true;
***************
*** 6706,6711 **** ATPrepAlterColumnType(List **wqueue,
--- 6708,6716 ----
  		/* Fix collations after all else */
  		assign_expr_collations(pstate, transform);
  
+ 		/* Plan the expr now so we can accurately assess the need to rewrite. */
+ 		transform = (Node *) expression_planner((Expr *) transform);
+ 
  		/*
  		 * Add a work queue item to make ATRewriteTable update the column
  		 * contents.
diff --git a/src/backend/optimizer/utilindex 2914c39..e9fb750 100644
*** a/src/backend/optimizer/util/clauses.c
--- b/src/backend/optimizer/util/clauses.c
***************
*** 106,114 **** static List *simplify_and_arguments(List *args,
  					   eval_const_expressions_context *context,
  					   bool *haveNull, bool *forceFalse);
  static Node *simplify_boolean_equality(Oid opno, List *args);
! static Expr *simplify_function(Oid funcid,
! 				  Oid result_type, int32 result_typmod,
! 				  Oid result_collid, Oid input_collid, List **args,
  				  bool has_named_args,
  				  bool allow_inline,
  				  eval_const_expressions_context *context);
--- 106,114 ----
  					   eval_const_expressions_context *context,
  					   bool *haveNull, bool *forceFalse);
  static Node *simplify_boolean_equality(Oid opno, List *args);
! static Expr *simplify_function(Expr *oldexpr, Oid funcid,
! 				  Oid result_type, int32 result_typmod, Oid result_collid,
! 				  Oid input_collid, List **args,
  				  bool has_named_args,
  				  bool allow_inline,
  				  eval_const_expressions_context *context);
***************
*** 2223,2229 **** eval_const_expressions_mutator(Node *node,
  		 * FuncExpr, but not when the node is recognizably a length coercion;
  		 * we want to preserve the typmod in the eventual Const if so.
  		 */
! 		simple = simplify_function(expr->funcid,
  								   expr->funcresulttype, exprTypmod(node),
  								   expr->funccollid,
  								   expr->inputcollid,
--- 2223,2230 ----
  		 * FuncExpr, but not when the node is recognizably a length coercion;
  		 * we want to preserve the typmod in the eventual Const if so.
  		 */
! 		simple = simplify_function((Expr *) expr,
! 								   expr->funcid,
  								   expr->funcresulttype, exprTypmod(node),
  								   expr->funccollid,
  								   expr->inputcollid,
***************
*** 2275,2281 **** eval_const_expressions_mutator(Node *node,
  		 * Code for op/func reduction is pretty bulky, so split it out as a
  		 * separate function.
  		 */
! 		simple = simplify_function(expr->opfuncid,
  								   expr->opresulttype, -1,
  								   expr->opcollid,
  								   expr->inputcollid,
--- 2276,2283 ----
  		 * Code for op/func reduction is pretty bulky, so split it out as a
  		 * separate function.
  		 */
! 		simple = simplify_function((Expr *) expr,
! 								   expr->opfuncid,
  								   expr->opresulttype, -1,
  								   expr->opcollid,
  								   expr->inputcollid,
***************
*** 2372,2378 **** eval_const_expressions_mutator(Node *node,
  			 * Code for op/func reduction is pretty bulky, so split it out as
  			 * a separate function.
  			 */
! 			simple = simplify_function(expr->opfuncid,
  									   expr->opresulttype, -1,
  									   expr->opcollid,
  									   expr->inputcollid,
--- 2374,2381 ----
  			 * Code for op/func reduction is pretty bulky, so split it out as
  			 * a separate function.
  			 */
! 			simple = simplify_function((Expr *) expr,
! 									   expr->opfuncid,
  									   expr->opresulttype, -1,
  									   expr->opcollid,
  									   expr->inputcollid,
***************
*** 2561,2567 **** eval_const_expressions_mutator(Node *node,
  		getTypeOutputInfo(exprType((Node *) arg), &outfunc, &outtypisvarlena);
  		getTypeInputInfo(expr->resulttype, &infunc, &intypioparam);
  
! 		simple = simplify_function(outfunc,
  								   CSTRINGOID, -1,
  								   InvalidOid,
  								   InvalidOid,
--- 2564,2571 ----
  		getTypeOutputInfo(exprType((Node *) arg), &outfunc, &outtypisvarlena);
  		getTypeInputInfo(expr->resulttype, &infunc, &intypioparam);
  
! 		simple = simplify_function(NULL,
! 								   outfunc,
  								   CSTRINGOID, -1,
  								   InvalidOid,
  								   InvalidOid,
***************
*** 2581,2587 **** eval_const_expressions_mutator(Node *node,
  									  Int32GetDatum(-1),
  									  false, true));
  
! 			simple = simplify_function(infunc,
  									   expr->resulttype, -1,
  									   expr->resultcollid,
  									   InvalidOid,
--- 2585,2592 ----
  									  Int32GetDatum(-1),
  									  false, true));
  
! 			simple = simplify_function(NULL,
! 									   infunc,
  									   expr->resulttype, -1,
  									   expr->resultcollid,
  									   InvalidOid,
***************
*** 3417,3427 **** simplify_boolean_equality(Oid opno, List *args)
   * Subroutine for eval_const_expressions: try to simplify a function call
   * (which might originally have been an operator; we don't care)
   *
!  * Inputs are the function OID, actual result type OID (which is needed for
!  * polymorphic functions), result typmod, result collation,
!  * the input collation to use for the function,
!  * the pre-simplified argument list, and some flags;
!  * also the context data for eval_const_expressions.
   *
   * Returns a simplified expression if successful, or NULL if cannot
   * simplify the function call.
--- 3422,3436 ----
   * Subroutine for eval_const_expressions: try to simplify a function call
   * (which might originally have been an operator; we don't care)
   *
!  * Inputs are the original expression (can be NULL), function OID, actual
!  * result type OID (which is needed for polymorphic functions), result typmod,
!  * result collation, the input collation to use for the function, the
!  * pre-simplified argument list, and some flags; also the context data for
!  * eval_const_expressions.  In common cases, several of the arguments could be
!  * derived from the original expression.  Sending them separately avoids
!  * duplicating NodeTag-specific knowledge, and it's necessary for CoerceViaIO.
!  * A NULL original expression disables use of transform functions while
!  * retaining all other behaviors.
   *
   * Returns a simplified expression if successful, or NULL if cannot
   * simplify the function call.
***************
*** 3433,3446 **** simplify_boolean_equality(Oid opno, List *args)
   * pass-by-reference, and it may get modified even if simplification fails.
   */
  static Expr *
! simplify_function(Oid funcid, Oid result_type, int32 result_typmod,
! 				  Oid result_collid, Oid input_collid, List **args,
  				  bool has_named_args,
  				  bool allow_inline,
  				  eval_const_expressions_context *context)
  {
  	HeapTuple	func_tuple;
  	Expr	   *newexpr;
  
  	/*
  	 * We have two strategies for simplification: either execute the function
--- 3442,3457 ----
   * pass-by-reference, and it may get modified even if simplification fails.
   */
  static Expr *
! simplify_function(Expr *oldexpr, Oid funcid,
! 				  Oid result_type, int32 result_typmod, Oid result_collid,
! 				  Oid input_collid, List **args,
  				  bool has_named_args,
  				  bool allow_inline,
  				  eval_const_expressions_context *context)
  {
  	HeapTuple	func_tuple;
  	Expr	   *newexpr;
+ 	Oid			transform;
  
  	/*
  	 * We have two strategies for simplification: either execute the function
***************
*** 3468,3473 **** simplify_function(Oid funcid, Oid result_type, int32 result_typmod,
--- 3479,3489 ----
  								result_collid, input_collid, *args,
  								func_tuple, context);
  
+ 	transform = ((Form_pg_proc) GETSTRUCT(func_tuple))->protransform;
+ 	if (!newexpr && OidIsValid(transform) && oldexpr)
+ 		newexpr = (Expr *) DatumGetPointer(OidFunctionCall1(transform,
+ 															PointerGetDatum(oldexpr)));
+ 
  	if (!newexpr && allow_inline)
  		newexpr = inline_function(funcid, result_type, result_collid,
  								  input_collid, *args,
diff --git a/src/backend/parser/parse_clausindex c5fe6b6..a8549e0 100644
*** a/src/backend/parser/parse_clause.c
--- b/src/backend/parser/parse_clause.c
***************
*** 2278,2280 **** transformFrameOffset(ParseState *pstate, int frameOptions, Node *clause)
--- 2278,2302 ----
  
  	return node;
  }
+ 
+ /*
+  * relabel_to_typmod
+  *		Add a RelabelType node that changes just the typmod, and remove all
+  *		now-superfluous RelabelType nodes beneath it.
+  */
+ Node *
+ relabel_to_typmod(Node *expr, int32 typmod)
+ {
+ 	Oid			type = exprType(expr);
+ 	Oid			coll = exprCollation(expr);
+ 
+ 	/*
+ 	 * Strip any existing RelabelType, then add one. This is to preserve the
+ 	 * invariant of no redundant RelabelTypes.
+ 	 */
+ 	while (IsA(expr, RelabelType))
+ 		expr = (Node *) ((RelabelType *) expr)->arg;
+ 
+ 	return (Node *) makeRelabelType((Expr *) expr, type, typmod, coll,
+ 									COERCE_DONTCARE);
+ }
diff --git a/src/backend/utils/adt/varchindex 1c0ef92..efa483d 100644
*** a/src/backend/utils/adt/varchar.c
--- b/src/backend/utils/adt/varchar.c
***************
*** 18,23 ****
--- 18,25 ----
  #include "access/hash.h"
  #include "access/tuptoaster.h"
  #include "libpq/pqformat.h"
+ #include "nodes/nodeFuncs.h"
+ #include "parser/parse_clause.h"
  #include "utils/array.h"
  #include "utils/builtins.h"
  #include "mb/pg_wchar.h"
***************
*** 549,554 **** varcharsend(PG_FUNCTION_ARGS)
--- 551,588 ----
  
  
  /*
+  * Flatten calls to our length coercion function that leave the new maximum
+  * length >= the previous maximum length.  We ignore the isExplicit argument,
+  * which only affects truncation.
+  */
+ Datum
+ varchar_transform(PG_FUNCTION_ARGS)
+ {
+ 	FuncExpr   *expr = (FuncExpr *) PG_GETARG_POINTER(0);
+ 	Node	   *typmod;
+ 	Node	   *ret = NULL;
+ 
+ 	if (!IsA(expr, FuncExpr))
+ 		PG_RETURN_POINTER(ret);
+ 
+ 	Assert(list_length(expr->args) == 3);
+ 	typmod = lsecond(expr->args);
+ 
+ 	if (IsA(typmod, Const))
+ 	{
+ 		Node	   *source = linitial(expr->args);
+ 		int32		new_typmod = DatumGetInt32(((Const *) typmod)->constvalue);
+ 		int32		old_max = exprTypmod(source) - VARHDRSZ;
+ 		int32		new_max = new_typmod - VARHDRSZ;
+ 
+ 		if (new_max < 0 || (old_max >= 0 && old_max <= new_max))
+ 			ret = relabel_to_typmod(source, new_typmod);
+ 	}
+ 
+ 	PG_RETURN_POINTER(ret);
+ }
+ 
+ /*
   * Converts a VARCHAR type to the specified size.
   *
   * maxlen is the typmod, ie, declared length plus VARHDRSZ bytes.
diff --git a/src/include/catalog/pg_clindex ffcce3c..002ae6b 100644
*** a/src/include/catalog/pg_class.h
--- b/src/include/catalog/pg_class.h
***************
*** 134,140 **** DATA(insert OID = 1247 (  pg_type		PGNSP 71 0 PGUID 0 0 0 0 0 0 0 f f p r 29 0 t
  DESCR("");
  DATA(insert OID = 1249 (  pg_attribute	PGNSP 75 0 PGUID 0 0 0 0 0 0 0 f f p r 20 0 f f f f f 3 _null_ _null_ ));
  DESCR("");
! DATA(insert OID = 1255 (  pg_proc		PGNSP 81 0 PGUID 0 0 0 0 0 0 0 f f p r 25 0 t f f f f 3 _null_ _null_ ));
  DESCR("");
  DATA(insert OID = 1259 (  pg_class		PGNSP 83 0 PGUID 0 0 0 0 0 0 0 f f p r 26 0 t f f f f 3 _null_ _null_ ));
  DESCR("");
--- 134,140 ----
  DESCR("");
  DATA(insert OID = 1249 (  pg_attribute	PGNSP 75 0 PGUID 0 0 0 0 0 0 0 f f p r 20 0 f f f f f 3 _null_ _null_ ));
  DESCR("");
! DATA(insert OID = 1255 (  pg_proc		PGNSP 81 0 PGUID 0 0 0 0 0 0 0 f f p r 26 0 t f f f f 3 _null_ _null_ ));
  DESCR("");
  DATA(insert OID = 1259 (  pg_class		PGNSP 83 0 PGUID 0 0 0 0 0 0 0 f f p r 26 0 t f f f f 3 _null_ _null_ ));
  DESCR("");
diff --git a/src/include/catalog/pg_pindex a66ecf1..6980d3e 100644
*** a/src/include/catalog/pg_proc.h
--- b/src/include/catalog/pg_proc.h
***************
*** 42,47 **** CATALOG(pg_proc,1255) BKI_BOOTSTRAP BKI_ROWTYPE_OID(81) BKI_SCHEMA_MACRO
--- 42,48 ----
  	float4		procost;		/* estimated execution cost */
  	float4		prorows;		/* estimated # of rows out (if proretset) */
  	Oid			provariadic;	/* element type of variadic array, or 0 */
+ 	regproc		protransform;	/* transforms calls to it during planning */
  	bool		proisagg;		/* is it an aggregate? */
  	bool		proiswindow;	/* is it a window function? */
  	bool		prosecdef;		/* security definer */
***************
*** 76,82 **** typedef FormData_pg_proc *Form_pg_proc;
   *		compiler constants for pg_proc
   * ----------------
   */
! #define Natts_pg_proc					25
  #define Anum_pg_proc_proname			1
  #define Anum_pg_proc_pronamespace		2
  #define Anum_pg_proc_proowner			3
--- 77,83 ----
   *		compiler constants for pg_proc
   * ----------------
   */
! #define Natts_pg_proc					26
  #define Anum_pg_proc_proname			1
  #define Anum_pg_proc_pronamespace		2
  #define Anum_pg_proc_proowner			3
***************
*** 84,107 **** typedef FormData_pg_proc *Form_pg_proc;
  #define Anum_pg_proc_procost			5
  #define Anum_pg_proc_prorows			6
  #define Anum_pg_proc_provariadic		7
! #define Anum_pg_proc_proisagg			8
! #define Anum_pg_proc_proiswindow		9
! #define Anum_pg_proc_prosecdef			10
! #define Anum_pg_proc_proisstrict		11
! #define Anum_pg_proc_proretset			12
! #define Anum_pg_proc_provolatile		13
! #define Anum_pg_proc_pronargs			14
! #define Anum_pg_proc_pronargdefaults	15
! #define Anum_pg_proc_prorettype			16
! #define Anum_pg_proc_proargtypes		17
! #define Anum_pg_proc_proallargtypes		18
! #define Anum_pg_proc_proargmodes		19
! #define Anum_pg_proc_proargnames		20
! #define Anum_pg_proc_proargdefaults		21
! #define Anum_pg_proc_prosrc				22
! #define Anum_pg_proc_probin				23
! #define Anum_pg_proc_proconfig			24
! #define Anum_pg_proc_proacl				25
  
  /* ----------------
   *		initial contents of pg_proc
--- 85,109 ----
  #define Anum_pg_proc_procost			5
  #define Anum_pg_proc_prorows			6
  #define Anum_pg_proc_provariadic		7
! #define Anum_pg_proc_protransform		8
! #define Anum_pg_proc_proisagg			9
! #define Anum_pg_proc_proiswindow		10
! #define Anum_pg_proc_prosecdef			11
! #define Anum_pg_proc_proisstrict		12
! #define Anum_pg_proc_proretset			13
! #define Anum_pg_proc_provolatile		14
! #define Anum_pg_proc_pronargs			15
! #define Anum_pg_proc_pronargdefaults	16
! #define Anum_pg_proc_prorettype			17
! #define Anum_pg_proc_proargtypes		18
! #define Anum_pg_proc_proallargtypes		19
! #define Anum_pg_proc_proargmodes		20
! #define Anum_pg_proc_proargnames		21
! #define Anum_pg_proc_proargdefaults		22
! #define Anum_pg_proc_prosrc				23
! #define Anum_pg_proc_probin				24
! #define Anum_pg_proc_proconfig			25
! #define Anum_pg_proc_proacl				26
  
  /* ----------------
   *		initial contents of pg_proc
***************
*** 743,749 **** DATA(insert OID = 659 (  namene			   PGNSP PGUID 12 1 0 0 0 f f f t f i 2 0 16 "
  
  DATA(insert OID = 668 (  bpchar			   PGNSP PGUID 12 1 0 0 0 f f f t f i 3 0 1042 "1042 23 16" _null_ _null_ _null_ _null_ bpchar _null_ _null_ _null_ ));
  DESCR("adjust char() to typmod length");
! DATA(insert OID = 669 (  varchar		   PGNSP PGUID 12 1 0 0 0 f f f t f i 3 0 1043 "1043 23 16" _null_ _null_ _null_ _null_ varchar _null_ _null_ _null_ ));
  DESCR("adjust varchar() to typmod length");
  
  DATA(insert OID = 676 (  mktinterval	   PGNSP PGUID 12 1 0 0 0 f f f t f i 2 0 704 "702 702" _null_ _null_ _null_ _null_ mktinterval _null_ _null_ _null_ ));
--- 745,753 ----
  
  DATA(insert OID = 668 (  bpchar			   PGNSP PGUID 12 1 0 0 0 f f f t f i 3 0 1042 "1042 23 16" _null_ _null_ _null_ _null_ bpchar _null_ _null_ _null_ ));
  DESCR("adjust char() to typmod length");
! DATA(insert OID = 3097 ( varchar_transform PGNSP PGUID 12 1 0 0 0 f f f t f i 1 0 2281 "2281" _null_ _null_ _null_ _null_ varchar_transform _null_ _null_ _null_ ));
! DESCR("transform a varchar length coercion");
! DATA(insert OID = 669 (  varchar		   PGNSP PGUID 12 1 0 0 3097 f f f t f i 3 0 1043 "1043 23 16" _null_ _null_ _null_ _null_ varchar _null_ _null_ _null_ ));
  DESCR("adjust varchar() to typmod length");
  
  DATA(insert OID = 676 (  mktinterval	   PGNSP PGUID 12 1 0 0 0 f f f t f i 2 0 704 "702 702" _null_ _null_ _null_ _null_ mktinterval _null_ _null_ _null_ ));
diff --git a/src/include/parser/parsindex 09ef450..f5ce2ff 100644
*** a/src/include/parser/parse_clause.h
--- b/src/include/parser/parse_clause.h
***************
*** 44,47 **** extern List *transformDistinctOnClause(ParseState *pstate, List *distinctlist,
--- 44,49 ----
  extern Index assignSortGroupRef(TargetEntry *tle, List *tlist);
  extern bool targetIsInSortList(TargetEntry *tle, Oid sortop, List *sortList);
  
+ extern Node *relabel_to_typmod(Node *expr, int32 typmod);
+ 
  #endif   /* PARSE_CLAUSE_H */
diff --git a/src/include/utils/builtins.index 14215db..8a1c82e 100644
*** a/src/include/utils/builtins.h
--- b/src/include/utils/builtins.h
***************
*** 684,689 **** extern Datum varcharrecv(PG_FUNCTION_ARGS);
--- 684,690 ----
  extern Datum varcharsend(PG_FUNCTION_ARGS);
  extern Datum varchartypmodin(PG_FUNCTION_ARGS);
  extern Datum varchartypmodout(PG_FUNCTION_ARGS);
+ extern Datum varchar_transform(PG_FUNCTION_ARGS);
  extern Datum varchar(PG_FUNCTION_ARGS);
  
  /* varlena.c */
diff --git a/src/test/regress/expecindex 3326f93..1b5a878 100644