format_with_width.diff

*** a/src/backend/utils/adt/varlena.c
--- b/src/backend/utils/adt/varlena.c
***************
*** 77,83 **** static bytea *bytea_substring(Datum str,
  static bytea *bytea_overlay(bytea *t1, bytea *t2, int sp, int sl);
  static StringInfo makeStringAggState(FunctionCallInfo fcinfo);
  void text_format_string_conversion(StringInfo buf, char conversion,
! 							  Oid typid, Datum value, bool isNull);
  
  static Datum text_to_array_internal(PG_FUNCTION_ARGS);
  static text *array_to_text_internal(FunctionCallInfo fcinfo, ArrayType *v,
--- 77,84 ----
  static bytea *bytea_overlay(bytea *t1, bytea *t2, int sp, int sl);
  static StringInfo makeStringAggState(FunctionCallInfo fcinfo);
  void text_format_string_conversion(StringInfo buf, char conversion,
! 							  Oid typid, Datum value, bool isNull,
! 							  bool with_width, int width);
  
  static Datum text_to_array_internal(PG_FUNCTION_ARGS);
  static text *array_to_text_internal(FunctionCallInfo fcinfo, ArrayType *v,
***************
*** 3947,3952 **** text_reverse(PG_FUNCTION_ARGS)
--- 3948,4028 ----
  }
  
  /*
+  * Returns ptr of first next char after digit. Raise error,
+  * when no digit is processed or when parsed values are not valid
+  */
+ static const char *
+ parse_digit_subformat(const char *start_ptr, const char *end_ptr, int *value)
+ {
+ 	const char *cp = start_ptr;
+ 	bool	minus = false;
+ 	int	inum = 0;
+ 
+ 	/* continue, only when start_ptr is less than end_ptr */
+ 	if (cp < end_ptr)
+ 	{
+ 		if (*cp == '-')
+ 		{
+ 			minus = true;
+ 			start_ptr = cp++;
+ 		}
+ 
+ 		while (cp < end_ptr)
+ 		{
+ 			if (*cp >= '0' && *cp <= '9')
+ 			{
+ 				int	newnum = inum * 10 + (*cp - '0');
+ 
+ 				if (newnum / 10 != inum) /* overflow? */
+ 					ereport(ERROR,
+ 							(errcode(ERRCODE_NUMERIC_VALUE_OUT_OF_RANGE),
+ 							 errmsg("number is out of range")));
+ 				inum = newnum;
+ 				++cp;
+ 			}
+ 			else
+ 				break;
+ 		}
+ 
+ 		*value = minus ? - inum : inum;
+ 	}
+ 
+ 	/* digit cannot be last char */
+ 	if (cp >= end_ptr)
+ 		ereport(ERROR,
+ 				(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
+ 				 errmsg("unterminated conversion specifier")));
+ 
+ 	/* we call this routine, only when we expected number */
+ 	if (cp == start_ptr)
+ 		ereport(ERROR,
+ 				(errcode(ERRCODE_NUMERIC_VALUE_OUT_OF_RANGE),
+ 				 errmsg("invalid format, missing number")));
+ 
+ 	/* argument number must by greather than zero */
+ 	if (*cp == '$')
+ 	{
+ 		if (minus)
+ 			ereport(ERROR,
+ 					(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
+ 					 errmsg("conversion specifies argument less 0, but arguments are numbered from 1")));
+ 
+ 		if (inum == 0)
+ 			ereport(ERROR,
+ 					(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
+ 					 errmsg("conversion specifies argument 0, but arguments are numbered from 1")));
+ 
+ 		/* $ must not be last char */
+ 		if (cp + 1 >= end_ptr)
+ 			ereport(ERROR,
+ 					(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
+ 					 errmsg("unterminated conversion specifier")));
+ 	}
+ 
+ 	return cp;
+ }
+ 
+ /*
   * Returns a formated string
   */
  Datum
***************
*** 3958,3964 **** text_format(PG_FUNCTION_ARGS)
  	const char *start_ptr;
  	const char *end_ptr;
  	text	   *result;
! 	int			arg = 0;
  
  	/* When format string is null, returns null */
  	if (PG_ARGISNULL(0))
--- 4034,4040 ----
  	const char *start_ptr;
  	const char *end_ptr;
  	text	   *result;
! 	int			args = 0;
  
  	/* When format string is null, returns null */
  	if (PG_ARGISNULL(0))
***************
*** 3976,3981 **** text_format(PG_FUNCTION_ARGS)
--- 4052,4060 ----
  		Datum		value;
  		bool		isNull;
  		Oid			typid;
+ 		bool		with_width;
+ 		int			arg;
+ 		int			width;
  
  		/*
  		 * If it's not the start of a conversion specifier, just copy it to
***************
*** 4000,4013 **** text_format(PG_FUNCTION_ARGS)
  			continue;
  		}
  
! 		/*
! 		 * If the user hasn't specified an argument position, we just advance
! 		 * to the next one.  If they have, we must parse it.
! 		 */
! 		if (*cp < '0' || *cp > '9')
  		{
! 			++arg;
! 			if (arg <= 0)		/* overflow? */
  			{
  				/*
  				 * Should not happen, as you can't pass billions of arguments
--- 4079,4124 ----
  			continue;
  		}
  
! 		if (cp >= end_ptr)
! 			ereport(ERROR,
! 					(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
! 					 errmsg("unterminated conversion specifier")));
! 
! 		arg = 0;
! 		with_width = false;
! 
! 		if (*cp == '-' || (*cp >= '0' && *cp <= '9'))
  		{
! 			cp = parse_digit_subformat(cp, end_ptr, &arg);
! 
! 			if (*cp == '$')
! 			{
! 				++cp;
! 				if (*cp == '-' || (*cp >= '0' && *cp <= '9'))
! 				{
! 					cp = parse_digit_subformat(cp, end_ptr, &width);
! 
! 					if (*cp == '$')
! 						ereport(ERROR,
! 								(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
! 								 errmsg("argument number is defined yet")));
! 
! 					with_width = true;
! 				}
! 			}
! 			else
! 			{
! 				/* previous digits was a width not argument number */
! 				width = arg;
! 				arg = 0;
! 				with_width = true;
! 			}
! 		}
! 
! 		if (arg == 0)
! 		{
! 			++args;
! 			if (args <= 0)		/* overflow? */
  			{
  				/*
  				 * Should not happen, as you can't pass billions of arguments
***************
*** 4017,4064 **** text_format(PG_FUNCTION_ARGS)
  						(errcode(ERRCODE_NUMERIC_VALUE_OUT_OF_RANGE),
  						 errmsg("argument number is out of range")));
  			}
- 		}
- 		else
- 		{
- 			bool		unterminated = false;
- 
- 			/* Parse digit string. */
- 			arg = 0;
- 			do
- 			{
- 				int			newarg = arg * 10 + (*cp - '0');
- 
- 				if (newarg / 10 != arg) /* overflow? */
- 					ereport(ERROR,
- 							(errcode(ERRCODE_NUMERIC_VALUE_OUT_OF_RANGE),
- 							 errmsg("argument number is out of range")));
- 				arg = newarg;
- 				++cp;
- 			} while (cp < end_ptr && *cp >= '0' && *cp <= '9');
- 
- 			/*
- 			 * If we ran off the end, or if there's not a $ next, or if the $
- 			 * is the last character, the conversion specifier is improperly
- 			 * terminated.
- 			 */
- 			if (cp == end_ptr || *cp != '$')
- 				unterminated = true;
- 			else
- 			{
- 				++cp;
- 				if (cp == end_ptr)
- 					unterminated = true;
- 			}
- 			if (unterminated)
- 				ereport(ERROR,
- 						(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
- 						 errmsg("unterminated conversion specifier")));
  
! 			/* There's no argument 0. */
! 			if (arg == 0)
! 				ereport(ERROR,
! 						(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
! 						 errmsg("conversion specifies argument 0, but arguments are numbered from 1")));
  		}
  
  		/* Not enough arguments?  Deduct 1 to avoid counting format string. */
--- 4128,4135 ----
  						(errcode(ERRCODE_NUMERIC_VALUE_OUT_OF_RANGE),
  						 errmsg("argument number is out of range")));
  			}
  
! 			arg = args;
  		}
  
  		/* Not enough arguments?  Deduct 1 to avoid counting format string. */
***************
*** 4081,4087 **** text_format(PG_FUNCTION_ARGS)
  			case 's':
  			case 'I':
  			case 'L':
! 				text_format_string_conversion(&str, *cp, typid, value, isNull);
  				break;
  			default:
  				ereport(ERROR,
--- 4152,4159 ----
  			case 's':
  			case 'I':
  			case 'L':
! 				text_format_string_conversion(&str, *cp, typid, value, isNull,
! 							      with_width, width);
  				break;
  			default:
  				ereport(ERROR,
***************
*** 4101,4112 **** text_format(PG_FUNCTION_ARGS)
  /* Format a %s, %I, or %L conversion. */
  void
  text_format_string_conversion(StringInfo buf, char conversion,
! 							  Oid typid, Datum value, bool isNull)
  {
  	Oid			typOutput;
  	bool		typIsVarlena;
  	char	   *str;
  
  	/* Handle NULL arguments before trying to stringify the value. */
  	if (isNull)
  	{
--- 4173,4190 ----
  /* Format a %s, %I, or %L conversion. */
  void
  text_format_string_conversion(StringInfo buf, char conversion,
! 							  Oid typid, Datum value, bool isNull,
! 							  bool with_width, int width)
  {
  	Oid			typOutput;
  	bool		typIsVarlena;
  	char	   *str;
  
+ 	if (conversion != 's' && with_width)
+ 		ereport(ERROR,
+ 				(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
+ 				 errmsg("width's specification is available only for %%s specifier")));
+ 
  	/* Handle NULL arguments before trying to stringify the value. */
  	if (isNull)
  	{
***************
*** 4138,4144 **** text_format_string_conversion(StringInfo buf, char conversion,
  		pfree(qstr);
  	}
  	else
! 		appendStringInfoString(buf, str);
  
  	/* Cleanup. */
  	pfree(str);
--- 4216,4245 ----
  		pfree(qstr);
  	}
  	else
! 	{
! 		/* fast path */
! 		if (!with_width)
! 			appendStringInfoString(buf, str);
! 		else
! 		{
! 			int	len = pg_mbstrlen(str);
! 
! 			if (width < 0)
! 			{
! 				/* allign to left */
! 				appendStringInfoString(buf, str);
! 				if (len < (-width))
! 					appendStringInfoSpaces(buf, - (len + width));
! 			}
! 			else
! 			{
! 				/* allign to right */
! 				if (len < width)
! 					appendStringInfoSpaces(buf, width - len);
! 				appendStringInfoString(buf, str);
! 			}
! 		}
! 	}
  
  	/* Cleanup. */
  	pfree(str);
*** a/src/test/regress/expected/text.out
--- b/src/test/regress/expected/text.out
***************
*** 223,233 **** ERROR:  too few arguments for format
  select format('%1$s %13$s', 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12);
  ERROR:  too few arguments for format
  select format('%1s', 1);
! ERROR:  unterminated conversion specifier
  select format('%1$', 1);
  ERROR:  unterminated conversion specifier
  select format('%1$1', 1);
! ERROR:  unrecognized conversion specifier "1"
  --checkk mix of positional and ordered placeholders
  select format('Hello %s %1$s %s', 'World', 'Hello again');
              format             
--- 223,237 ----
  select format('%1$s %13$s', 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12);
  ERROR:  too few arguments for format
  select format('%1s', 1);
!  format 
! --------
!  1
! (1 row)
! 
  select format('%1$', 1);
  ERROR:  unterminated conversion specifier
  select format('%1$1', 1);
! ERROR:  unterminated conversion specifier
  --checkk mix of positional and ordered placeholders
  select format('Hello %s %1$s %s', 'World', 'Hello again');
              format             
***************
*** 241,243 **** select format('Hello %s %s, %2$s %2$s', 'World', 'Hello again');
--- 245,263 ----
   Hello World Hello again, Hello again Hello again
  (1 row)
  
+ select format('%s %2$s %s', 'Hello', 'World');
+       format       
+ -------------------
+  Hello World World
+ (1 row)
+ 
+ --check width
+ select format('||%4s|| ||%-4s||', 'ab', 'ab');
+       format       
+ -------------------
+  ||  ab|| ||ab  ||
+ (1 row)
+ 
+ --should fail
+ select format('INSERT INTO %10I VALUES(%L,%L)', 'mytab', 10, 'Hello');
+ ERROR:  width's specification is available only for %s specifier
*** a/src/test/regress/sql/text.sql
--- b/src/test/regress/sql/text.sql
***************
*** 76,78 **** select format('%1$1', 1);
--- 76,86 ----
  --checkk mix of positional and ordered placeholders
  select format('Hello %s %1$s %s', 'World', 'Hello again');
  select format('Hello %s %s, %2$s %2$s', 'World', 'Hello again');
+ 
+ select format('%s %2$s %s', 'Hello', 'World');
+ 
+ --check width
+ select format('||%4s|| ||%-4s||', 'ab', 'ab');
+ 
+ --should fail
+ select format('INSERT INTO %10I VALUES(%L,%L)', 'mytab', 10, 'Hello');