Get rid of backtracking in jsonpath_scan.l

Non-backtracking flex parsers work faster than backtracking ones.  So, this
commit gets rid of backtracking in jsonpath_scan.l.  That required explicit
handling of some cases as well as manual backtracking for some cases.  More
regression tests for numerics are added.

Discussion: https://mail.google.com/mail/u/0?ik=a20b091faa&view=om&permmsgid=msg-f%3A1628425344167939063
Author: John Naylor, Nikita Gluknov, Alexander Korotkov

src/backend/utils/adt/Makefile

@@ -34,6 +34,7 @@ OBJS = acl.o amutils.o arrayfuncs.o array_expanded.o array_selfuncs.o \
 	windowfuncs.o xid.o xml.o
 
 jsonpath_scan.c: FLEXFLAGS = -CF -p -p
+jsonpath_scan.c: FLEX_NO_BACKUP=yes
 
 # Force these dependencies to be known even without dependency info built:
 jsonpath_gram.o: jsonpath_scan.c

src/backend/utils/adt/jsonpath_scan.l

@@ -31,7 +31,7 @@ static void addstring(bool init, char *s, int l);
 static void addchar(bool init, char s);
 static enum yytokentype checkKeyword(void);
 static void parseUnicode(char *s, int l);
-static void parseHexChars(char *s, int l);
+static void parseHexChar(char *s);
 
 /* Avoid exit() on fatal scanner errors (a bit ugly -- see yy_fatal_error) */
 #undef fprintf
@@ -78,9 +78,20 @@ fprintf_to_ereport(const char *fmt, const char *msg)
 special		 [\?\%\$\.\[\]\{\}\(\)\|\&\!\=\<\>\@\#\,\*:\-\+\/]
 any			[^\?\%\$\.\[\]\{\}\(\)\|\&\!\=\<\>\@\#\,\*:\-\+\/\\\"\' \t\n\r\f]
 blank		[ \t\n\r\f]
+
+digit		[0-9]
+integer		{digit}+
+decimal		{digit}*\.{digit}+
+decimalfail	{digit}+\.
+real		({integer}|{decimal})[Ee][-+]?{digit}+
+realfail1	({integer}|{decimal})[Ee]
+realfail2	({integer}|{decimal})[Ee][-+]
+
 hex_dig		[0-9A-Fa-f]
 unicode		\\u({hex_dig}{4}|\{{hex_dig}{1,6}\})
+unicodefail	\\u({hex_dig}{0,3}|\{{hex_dig}{0,6})
 hex_char	\\x{hex_dig}{2}
+hex_fail	\\x{hex_dig}{0,1}
 
 %%
 
@@ -129,11 +140,17 @@ hex_char	\\x{hex_dig}{2}
 
 <xnq,xq,xvq,xsq>{unicode}+		{ parseUnicode(yytext, yyleng); }
 
-<xnq,xq,xvq,xsq>{hex_char}+		{ parseHexChars(yytext, yyleng); }
+<xnq,xq,xvq,xsq>{hex_char}		{ parseHexChar(yytext); }
+
+<xnq,xq,xvq,xsq>{unicode}*{unicodefail}	{ yyerror(NULL, "Unicode sequence is invalid"); }
 
-<xnq,xq,xvq,xsq>\\x				{ yyerror(NULL, "Hex character sequence is invalid"); }
+<xnq,xq,xvq,xsq>{hex_fail}		{ yyerror(NULL, "Hex character sequence is invalid"); }
 
-<xnq,xq,xvq,xsq>\\u				{ yyerror(NULL, "Unicode sequence is invalid"); }
+<xnq,xq,xvq,xsq>{unicode}+\\	{
+									/* throw back the \\, and treat as unicode */
+									yyless(yyleng - 1);
+									parseUnicode(yytext, yyleng);
+								}
 
 <xnq,xq,xvq,xsq>\\.				{ yyerror(NULL, "Escape sequence is invalid"); }
 
@@ -214,34 +231,38 @@ hex_char	\\x{hex_dig}{2}
 									BEGIN xc;
 								}
 
-[0-9]+(\.[0-9]+)?[eE][+-]?[0-9]+ { /* float */
+{real}							{
 									addstring(true, yytext, yyleng);
 									addchar(false, '\0');
 									yylval->str = scanstring;
 									return NUMERIC_P;
 								}
 
-\.[0-9]+[eE][+-]?[0-9]+			{ /* float */
+{decimal}						{
 									addstring(true, yytext, yyleng);
 									addchar(false, '\0');
 									yylval->str = scanstring;
 									return NUMERIC_P;
 								}
 
-([0-9]+)?\.[0-9]+				{
+{integer}						{
 									addstring(true, yytext, yyleng);
 									addchar(false, '\0');
 									yylval->str = scanstring;
-									return NUMERIC_P;
+									return INT_P;
 								}
 
-[0-9]+							{
+{decimalfail}					{
+									/* throw back the ., and treat as integer */
+									yyless(yyleng - 1);
 									addstring(true, yytext, yyleng);
 									addchar(false, '\0');
 									yylval->str = scanstring;
 									return INT_P;
 								}
 
+({realfail1}|{realfail2})		{ yyerror(NULL, "Floating point number is invalid"); }
+
 {any}+							{
 									addstring(true, yytext, yyleng);
 									BEGIN xnq;
@@ -571,7 +592,7 @@ addUnicode(int ch, int *hi_surrogate)
 static void
 parseUnicode(char *s, int l)
 {
-	int			i;
+	int			i = 2;
 	int			hi_surrogate = -1;
 
 	for (i = 2; i < l; i += 2)	/* skip '\u' */
@@ -606,19 +627,12 @@ parseUnicode(char *s, int l)
 
 /* Parse sequence of hex-encoded characters */
 static void
-parseHexChars(char *s, int l)
+parseHexChar(char *s)
 {
-	int i;
+	int			ch = (hexval(s[2]) << 4) |
-
+					  hexval(s[3]);
-	Assert(l % 4 /* \xXX */ == 0);
-
-	for (i = 0; i < l / 4; i++)
-	{
-		int			ch = (hexval(s[i * 4 + 2]) << 4) |
-						  hexval(s[i * 4 + 3]);
 
-		addUnicodeChar(ch);
+	addUnicodeChar(ch);
-	}
 }
 
 /*

src/test/regress/expected/jsonb_jsonpath.out

@@ -1297,7 +1297,7 @@ select jsonb_path_query('null', 'true.type()');
  "boolean"
 (1 row)
 
-select jsonb_path_query('null', '123.type()');
+select jsonb_path_query('null', '(123).type()');
  jsonb_path_query 
 ------------------
  "number"

src/test/regress/expected/jsonpath.out

@@ -365,6 +365,18 @@ select '1.type()'::jsonpath;
  1.type()
 (1 row)
 
+select '(1).type()'::jsonpath;
+ jsonpath 
+----------
+ 1.type()
+(1 row)
+
+select '1.2.type()'::jsonpath;
+  jsonpath  
+------------
+ 1.2.type()
+(1 row)
+
 select '"aaa".type()'::jsonpath;
    jsonpath   
 --------------
@@ -804,3 +816,159 @@ select '$ ? (@.a < +10.1e+1)'::jsonpath;
  $?(@."a" < 101)
 (1 row)
 
+select '0'::jsonpath;
+ jsonpath 
+----------
+ 0
+(1 row)
+
+select '00'::jsonpath;
+ jsonpath 
+----------
+ 0
+(1 row)
+
+select '0.0'::jsonpath;
+ jsonpath 
+----------
+ 0.0
+(1 row)
+
+select '0.000'::jsonpath;
+ jsonpath 
+----------
+ 0.000
+(1 row)
+
+select '0.000e1'::jsonpath;
+ jsonpath 
+----------
+ 0.00
+(1 row)
+
+select '0.000e2'::jsonpath;
+ jsonpath 
+----------
+ 0.0
+(1 row)
+
+select '0.000e3'::jsonpath;
+ jsonpath 
+----------
+ 0
+(1 row)
+
+select '0.0010'::jsonpath;
+ jsonpath 
+----------
+ 0.0010
+(1 row)
+
+select '0.0010e-1'::jsonpath;
+ jsonpath 
+----------
+ 0.00010
+(1 row)
+
+select '0.0010e+1'::jsonpath;
+ jsonpath 
+----------
+ 0.010
+(1 row)
+
+select '0.0010e+2'::jsonpath;
+ jsonpath 
+----------
+ 0.10
+(1 row)
+
+select '1e'::jsonpath;
+ERROR:  bad jsonpath representation
+LINE 1: select '1e'::jsonpath;
+               ^
+DETAIL:  Floating point number is invalid at or near "1e"
+select '1.e'::jsonpath;
+ jsonpath 
+----------
+ 1."e"
+(1 row)
+
+select '1.2e'::jsonpath;
+ERROR:  bad jsonpath representation
+LINE 1: select '1.2e'::jsonpath;
+               ^
+DETAIL:  Floating point number is invalid at or near "1.2e"
+select '1.2.e'::jsonpath;
+ jsonpath 
+----------
+ 1.2."e"
+(1 row)
+
+select '(1.2).e'::jsonpath;
+ jsonpath 
+----------
+ 1.2."e"
+(1 row)
+
+select '1e3'::jsonpath;
+ jsonpath 
+----------
+ 1000
+(1 row)
+
+select '1.e3'::jsonpath;
+ jsonpath 
+----------
+ 1."e3"
+(1 row)
+
+select '1.e3.e'::jsonpath;
+  jsonpath  
+------------
+ 1."e3"."e"
+(1 row)
+
+select '1.e3.e4'::jsonpath;
+  jsonpath   
+-------------
+ 1."e3"."e4"
+(1 row)
+
+select '1.2e3'::jsonpath;
+ jsonpath 
+----------
+ 1200
+(1 row)
+
+select '1.2.e3'::jsonpath;
+ jsonpath 
+----------
+ 1.2."e3"
+(1 row)
+
+select '(1.2).e3'::jsonpath;
+ jsonpath 
+----------
+ 1.2."e3"
+(1 row)
+
+select '1..e'::jsonpath;
+ERROR:  bad jsonpath representation
+LINE 1: select '1..e'::jsonpath;
+               ^
+DETAIL:  syntax error, unexpected '.' at or near "."
+select '1..e3'::jsonpath;
+ERROR:  bad jsonpath representation
+LINE 1: select '1..e3'::jsonpath;
+               ^
+DETAIL:  syntax error, unexpected '.' at or near "."
+select '(1.).e'::jsonpath;
+ERROR:  bad jsonpath representation
+LINE 1: select '(1.).e'::jsonpath;
+               ^
+DETAIL:  syntax error, unexpected ')' at or near ")"
+select '(1.).e3'::jsonpath;
+ERROR:  bad jsonpath representation
+LINE 1: select '(1.).e3'::jsonpath;
+               ^
+DETAIL:  syntax error, unexpected ')' at or near ")"

src/test/regress/expected/jsonpath_encoding.out

@@ -0,0 +1,249 @@
+-- encoding-sensitive tests for jsonpath
+-- checks for double-quoted values
+-- basic unicode input
+SELECT '"\u"'::jsonpath;		-- ERROR, incomplete escape
+ERROR:  bad jsonpath representation
+LINE 1: SELECT '"\u"'::jsonpath;
+               ^
+DETAIL:  Unicode sequence is invalid at or near "\u"
+SELECT '"\u00"'::jsonpath;		-- ERROR, incomplete escape
+ERROR:  bad jsonpath representation
+LINE 1: SELECT '"\u00"'::jsonpath;
+               ^
+DETAIL:  Unicode sequence is invalid at or near "\u00"
+SELECT '"\u000g"'::jsonpath;	-- ERROR, g is not a hex digit
+ERROR:  bad jsonpath representation
+LINE 1: SELECT '"\u000g"'::jsonpath;
+               ^
+DETAIL:  Unicode sequence is invalid at or near "\u000"
+SELECT '"\u0000"'::jsonpath;	-- OK, legal escape
+ERROR:  unsupported Unicode escape sequence
+LINE 1: SELECT '"\u0000"'::jsonpath;
+               ^
+DETAIL:  \u0000 cannot be converted to text.
+SELECT '"\uaBcD"'::jsonpath;	-- OK, uppercase and lower case both OK
+ jsonpath 
+----------
+ "ꯍ"
+(1 row)
+
+-- handling of unicode surrogate pairs
+select '"\ud83d\ude04\ud83d\udc36"'::jsonpath as correct_in_utf8;
+ correct_in_utf8 
+-----------------
+ "😄🐶"
+(1 row)
+
+select '"\ud83d\ud83d"'::jsonpath; -- 2 high surrogates in a row
+ERROR:  invalid input syntax for type jsonpath
+LINE 1: select '"\ud83d\ud83d"'::jsonpath;
+               ^
+DETAIL:  Unicode high surrogate must not follow a high surrogate.
+select '"\ude04\ud83d"'::jsonpath; -- surrogates in wrong order
+ERROR:  invalid input syntax for type jsonpath
+LINE 1: select '"\ude04\ud83d"'::jsonpath;
+               ^
+DETAIL:  Unicode low surrogate must follow a high surrogate.
+select '"\ud83dX"'::jsonpath; -- orphan high surrogate
+ERROR:  invalid input syntax for type jsonpath
+LINE 1: select '"\ud83dX"'::jsonpath;
+               ^
+DETAIL:  Unicode low surrogate must follow a high surrogate.
+select '"\ude04X"'::jsonpath; -- orphan low surrogate
+ERROR:  invalid input syntax for type jsonpath
+LINE 1: select '"\ude04X"'::jsonpath;
+               ^
+DETAIL:  Unicode low surrogate must follow a high surrogate.
+--handling of simple unicode escapes
+select '"the Copyright \u00a9 sign"'::jsonpath as correct_in_utf8;
+    correct_in_utf8     
+------------------------
+ "the Copyright © sign"
+(1 row)
+
+select '"dollar \u0024 character"'::jsonpath as correct_everywhere;
+  correct_everywhere  
+----------------------
+ "dollar $ character"
+(1 row)
+
+select '"dollar \\u0024 character"'::jsonpath as not_an_escape;
+       not_an_escape        
+----------------------------
+ "dollar \\u0024 character"
+(1 row)
+
+select '"null \u0000 escape"'::jsonpath as not_unescaped;
+ERROR:  unsupported Unicode escape sequence
+LINE 1: select '"null \u0000 escape"'::jsonpath as not_unescaped;
+               ^
+DETAIL:  \u0000 cannot be converted to text.
+select '"null \\u0000 escape"'::jsonpath as not_an_escape;
+     not_an_escape     
+-----------------------
+ "null \\u0000 escape"
+(1 row)
+
+-- checks for single-quoted values
+-- basic unicode input
+SELECT E'\'\u\''::jsonpath;		-- ERROR, incomplete escape
+ERROR:  invalid Unicode escape
+LINE 1: SELECT E'\'\u\''::jsonpath;
+               ^
+HINT:  Unicode escapes must be \uXXXX or \UXXXXXXXX.
+SELECT E'\'\u00\''::jsonpath;	-- ERROR, incomplete escape
+ERROR:  invalid Unicode escape
+LINE 1: SELECT E'\'\u00\''::jsonpath;
+               ^
+HINT:  Unicode escapes must be \uXXXX or \UXXXXXXXX.
+SELECT E'\'\u000g\''::jsonpath;	-- ERROR, g is not a hex digit
+ERROR:  invalid Unicode escape
+LINE 1: SELECT E'\'\u000g\''::jsonpath;
+               ^
+HINT:  Unicode escapes must be \uXXXX or \UXXXXXXXX.
+SELECT E'\'\u0000\''::jsonpath;	-- OK, legal escape
+ERROR:  invalid Unicode escape value at or near "E'\'\u0000"
+LINE 1: SELECT E'\'\u0000\''::jsonpath;
+               ^
+SELECT E'\'\uaBcD\''::jsonpath;	-- OK, uppercase and lower case both OK
+ jsonpath 
+----------
+ "ꯍ"
+(1 row)
+
+-- handling of unicode surrogate pairs
+select E'\'\ud83d\ude04\ud83d\udc36\''::jsonpath as correct_in_utf8;
+ correct_in_utf8 
+-----------------
+ "😄🐶"
+(1 row)
+
+select E'\'\ud83d\ud83d\''::jsonpath; -- 2 high surrogates in a row
+ERROR:  invalid Unicode surrogate pair at or near "E'\'\ud83d\ud83d"
+LINE 1: select E'\'\ud83d\ud83d\''::jsonpath;
+               ^
+select E'\'\ude04\ud83d\''::jsonpath; -- surrogates in wrong order
+ERROR:  invalid Unicode surrogate pair at or near "E'\'\ude04"
+LINE 1: select E'\'\ude04\ud83d\''::jsonpath;
+               ^
+select E'\'\ud83dX\''::jsonpath; -- orphan high surrogate
+ERROR:  invalid Unicode surrogate pair at or near "E'\'\ud83dX"
+LINE 1: select E'\'\ud83dX\''::jsonpath;
+               ^
+select E'\'\ude04X\''::jsonpath; -- orphan low surrogate
+ERROR:  invalid Unicode surrogate pair at or near "E'\'\ude04"
+LINE 1: select E'\'\ude04X\''::jsonpath;
+               ^
+--handling of simple unicode escapes
+select E'\'the Copyright \u00a9 sign\''::jsonpath as correct_in_utf8;
+    correct_in_utf8     
+------------------------
+ "the Copyright © sign"
+(1 row)
+
+select E'\'dollar \u0024 character\''::jsonpath as correct_everywhere;
+  correct_everywhere  
+----------------------
+ "dollar $ character"
+(1 row)
+
+select E'\'dollar \\u0024 character\''::jsonpath as not_an_escape;
+    not_an_escape     
+----------------------
+ "dollar $ character"
+(1 row)
+
+select E'\'null \u0000 escape\''::jsonpath as not_unescaped;
+ERROR:  invalid Unicode escape value at or near "E'\'null \u0000"
+LINE 1: select E'\'null \u0000 escape\''::jsonpath as not_unescaped;
+               ^
+select E'\'null \\u0000 escape\''::jsonpath as not_an_escape;
+ERROR:  unsupported Unicode escape sequence
+LINE 1: select E'\'null \\u0000 escape\''::jsonpath as not_an_escape...
+               ^
+DETAIL:  \u0000 cannot be converted to text.
+-- checks for quoted key names
+-- basic unicode input
+SELECT '$."\u"'::jsonpath;		-- ERROR, incomplete escape
+ERROR:  bad jsonpath representation
+LINE 1: SELECT '$."\u"'::jsonpath;
+               ^
+DETAIL:  Unicode sequence is invalid at or near "\u"
+SELECT '$."\u00"'::jsonpath;	-- ERROR, incomplete escape
+ERROR:  bad jsonpath representation
+LINE 1: SELECT '$."\u00"'::jsonpath;
+               ^
+DETAIL:  Unicode sequence is invalid at or near "\u00"
+SELECT '$."\u000g"'::jsonpath;	-- ERROR, g is not a hex digit
+ERROR:  bad jsonpath representation
+LINE 1: SELECT '$."\u000g"'::jsonpath;
+               ^
+DETAIL:  Unicode sequence is invalid at or near "\u000"
+SELECT '$."\u0000"'::jsonpath;	-- OK, legal escape
+ERROR:  unsupported Unicode escape sequence
+LINE 1: SELECT '$."\u0000"'::jsonpath;
+               ^
+DETAIL:  \u0000 cannot be converted to text.
+SELECT '$."\uaBcD"'::jsonpath;	-- OK, uppercase and lower case both OK
+ jsonpath 
+----------
+ $."ꯍ"
+(1 row)
+
+-- handling of unicode surrogate pairs
+select '$."\ud83d\ude04\ud83d\udc36"'::jsonpath as correct_in_utf8;
+ correct_in_utf8 
+-----------------
+ $."😄🐶"
+(1 row)
+
+select '$."\ud83d\ud83d"'::jsonpath; -- 2 high surrogates in a row
+ERROR:  invalid input syntax for type jsonpath
+LINE 1: select '$."\ud83d\ud83d"'::jsonpath;
+               ^
+DETAIL:  Unicode high surrogate must not follow a high surrogate.
+select '$."\ude04\ud83d"'::jsonpath; -- surrogates in wrong order
+ERROR:  invalid input syntax for type jsonpath
+LINE 1: select '$."\ude04\ud83d"'::jsonpath;
+               ^
+DETAIL:  Unicode low surrogate must follow a high surrogate.
+select '$."\ud83dX"'::jsonpath; -- orphan high surrogate
+ERROR:  invalid input syntax for type jsonpath
+LINE 1: select '$."\ud83dX"'::jsonpath;
+               ^
+DETAIL:  Unicode low surrogate must follow a high surrogate.
+select '$."\ude04X"'::jsonpath; -- orphan low surrogate
+ERROR:  invalid input syntax for type jsonpath
+LINE 1: select '$."\ude04X"'::jsonpath;
+               ^
+DETAIL:  Unicode low surrogate must follow a high surrogate.
+--handling of simple unicode escapes
+select '$."the Copyright \u00a9 sign"'::jsonpath as correct_in_utf8;
+     correct_in_utf8      
+--------------------------
+ $."the Copyright © sign"
+(1 row)
+
+select '$."dollar \u0024 character"'::jsonpath as correct_everywhere;
+   correct_everywhere   
+------------------------
+ $."dollar $ character"
+(1 row)
+
+select '$."dollar \\u0024 character"'::jsonpath as not_an_escape;
+        not_an_escape         
+------------------------------
+ $."dollar \\u0024 character"
+(1 row)
+
+select '$."null \u0000 escape"'::jsonpath as not_unescaped;
+ERROR:  unsupported Unicode escape sequence
+LINE 1: select '$."null \u0000 escape"'::jsonpath as not_unescaped;
+               ^
+DETAIL:  \u0000 cannot be converted to text.
+select '$."null \\u0000 escape"'::jsonpath as not_an_escape;
+      not_an_escape      
+-------------------------
+ $."null \\u0000 escape"
+(1 row)
+

src/test/regress/expected/jsonpath_encoding_1.out

@@ -0,0 +1,237 @@
+-- encoding-sensitive tests for jsonpath
+-- checks for double-quoted values
+-- basic unicode input
+SELECT '"\u"'::jsonpath;		-- ERROR, incomplete escape
+ERROR:  bad jsonpath representation
+LINE 1: SELECT '"\u"'::jsonpath;
+               ^
+DETAIL:  Unicode sequence is invalid at or near "\u"
+SELECT '"\u00"'::jsonpath;		-- ERROR, incomplete escape
+ERROR:  bad jsonpath representation
+LINE 1: SELECT '"\u00"'::jsonpath;
+               ^
+DETAIL:  Unicode sequence is invalid at or near "\u00"
+SELECT '"\u000g"'::jsonpath;	-- ERROR, g is not a hex digit
+ERROR:  bad jsonpath representation
+LINE 1: SELECT '"\u000g"'::jsonpath;
+               ^
+DETAIL:  Unicode sequence is invalid at or near "\u000"
+SELECT '"\u0000"'::jsonpath;	-- OK, legal escape
+ERROR:  unsupported Unicode escape sequence
+LINE 1: SELECT '"\u0000"'::jsonpath;
+               ^
+DETAIL:  \u0000 cannot be converted to text.
+SELECT '"\uaBcD"'::jsonpath;	-- OK, uppercase and lower case both OK
+ERROR:  invalid input syntax for type jsonpath
+LINE 1: SELECT '"\uaBcD"'::jsonpath;
+               ^
+DETAIL:  Unicode escape values cannot be used for code point values above 007F when the server encoding is not UTF8.
+-- handling of unicode surrogate pairs
+select '"\ud83d\ude04\ud83d\udc36"'::jsonpath as correct_in_utf8;
+ERROR:  invalid input syntax for type jsonpath
+LINE 1: select '"\ud83d\ude04\ud83d\udc36"'::jsonpath as correct_in_...
+               ^
+DETAIL:  Unicode escape values cannot be used for code point values above 007F when the server encoding is not UTF8.
+select '"\ud83d\ud83d"'::jsonpath; -- 2 high surrogates in a row
+ERROR:  invalid input syntax for type jsonpath
+LINE 1: select '"\ud83d\ud83d"'::jsonpath;
+               ^
+DETAIL:  Unicode high surrogate must not follow a high surrogate.
+select '"\ude04\ud83d"'::jsonpath; -- surrogates in wrong order
+ERROR:  invalid input syntax for type jsonpath
+LINE 1: select '"\ude04\ud83d"'::jsonpath;
+               ^
+DETAIL:  Unicode low surrogate must follow a high surrogate.
+select '"\ud83dX"'::jsonpath; -- orphan high surrogate
+ERROR:  invalid input syntax for type jsonpath
+LINE 1: select '"\ud83dX"'::jsonpath;
+               ^
+DETAIL:  Unicode low surrogate must follow a high surrogate.
+select '"\ude04X"'::jsonpath; -- orphan low surrogate
+ERROR:  invalid input syntax for type jsonpath
+LINE 1: select '"\ude04X"'::jsonpath;
+               ^
+DETAIL:  Unicode low surrogate must follow a high surrogate.
+--handling of simple unicode escapes
+select '"the Copyright \u00a9 sign"'::jsonpath as correct_in_utf8;
+ERROR:  invalid input syntax for type jsonpath
+LINE 1: select '"the Copyright \u00a9 sign"'::jsonpath as correct_in...
+               ^
+DETAIL:  Unicode escape values cannot be used for code point values above 007F when the server encoding is not UTF8.
+select '"dollar \u0024 character"'::jsonpath as correct_everywhere;
+  correct_everywhere  
+----------------------
+ "dollar $ character"
+(1 row)
+
+select '"dollar \\u0024 character"'::jsonpath as not_an_escape;
+       not_an_escape        
+----------------------------
+ "dollar \\u0024 character"
+(1 row)
+
+select '"null \u0000 escape"'::jsonpath as not_unescaped;
+ERROR:  unsupported Unicode escape sequence
+LINE 1: select '"null \u0000 escape"'::jsonpath as not_unescaped;
+               ^
+DETAIL:  \u0000 cannot be converted to text.
+select '"null \\u0000 escape"'::jsonpath as not_an_escape;
+     not_an_escape     
+-----------------------
+ "null \\u0000 escape"
+(1 row)
+
+-- checks for single-quoted values
+-- basic unicode input
+SELECT E'\'\u\''::jsonpath;		-- ERROR, incomplete escape
+ERROR:  invalid Unicode escape
+LINE 1: SELECT E'\'\u\''::jsonpath;
+               ^
+HINT:  Unicode escapes must be \uXXXX or \UXXXXXXXX.
+SELECT E'\'\u00\''::jsonpath;	-- ERROR, incomplete escape
+ERROR:  invalid Unicode escape
+LINE 1: SELECT E'\'\u00\''::jsonpath;
+               ^
+HINT:  Unicode escapes must be \uXXXX or \UXXXXXXXX.
+SELECT E'\'\u000g\''::jsonpath;	-- ERROR, g is not a hex digit
+ERROR:  invalid Unicode escape
+LINE 1: SELECT E'\'\u000g\''::jsonpath;
+               ^
+HINT:  Unicode escapes must be \uXXXX or \UXXXXXXXX.
+SELECT E'\'\u0000\''::jsonpath;	-- OK, legal escape
+ERROR:  invalid Unicode escape value at or near "E'\'\u0000"
+LINE 1: SELECT E'\'\u0000\''::jsonpath;
+               ^
+SELECT E'\'\uaBcD\''::jsonpath;	-- OK, uppercase and lower case both OK
+ERROR:  Unicode escape values cannot be used for code point values above 007F when the server encoding is not UTF8 at or near "E'\'\uaBcD"
+LINE 1: SELECT E'\'\uaBcD\''::jsonpath;
+               ^
+-- handling of unicode surrogate pairs
+select E'\'\ud83d\ude04\ud83d\udc36\''::jsonpath as correct_in_utf8;
+ERROR:  Unicode escape values cannot be used for code point values above 007F when the server encoding is not UTF8 at or near "E'\'\ud83d\ude04"
+LINE 1: select E'\'\ud83d\ude04\ud83d\udc36\''::jsonpath as correct_...
+               ^
+select E'\'\ud83d\ud83d\''::jsonpath; -- 2 high surrogates in a row
+ERROR:  invalid Unicode surrogate pair at or near "E'\'\ud83d\ud83d"
+LINE 1: select E'\'\ud83d\ud83d\''::jsonpath;
+               ^
+select E'\'\ude04\ud83d\''::jsonpath; -- surrogates in wrong order
+ERROR:  invalid Unicode surrogate pair at or near "E'\'\ude04"
+LINE 1: select E'\'\ude04\ud83d\''::jsonpath;
+               ^
+select E'\'\ud83dX\''::jsonpath; -- orphan high surrogate
+ERROR:  invalid Unicode surrogate pair at or near "E'\'\ud83dX"
+LINE 1: select E'\'\ud83dX\''::jsonpath;
+               ^
+select E'\'\ude04X\''::jsonpath; -- orphan low surrogate
+ERROR:  invalid Unicode surrogate pair at or near "E'\'\ude04"
+LINE 1: select E'\'\ude04X\''::jsonpath;
+               ^
+--handling of simple unicode escapes
+select E'\'the Copyright \u00a9 sign\''::jsonpath as correct_in_utf8;
+ERROR:  Unicode escape values cannot be used for code point values above 007F when the server encoding is not UTF8 at or near "E'\'the Copyright \u00a9"
+LINE 1: select E'\'the Copyright \u00a9 sign\''::jsonpath as correct...
+               ^
+select E'\'dollar \u0024 character\''::jsonpath as correct_everywhere;
+  correct_everywhere  
+----------------------
+ "dollar $ character"
+(1 row)
+
+select E'\'dollar \\u0024 character\''::jsonpath as not_an_escape;
+    not_an_escape     
+----------------------
+ "dollar $ character"
+(1 row)
+
+select E'\'null \u0000 escape\''::jsonpath as not_unescaped;
+ERROR:  invalid Unicode escape value at or near "E'\'null \u0000"
+LINE 1: select E'\'null \u0000 escape\''::jsonpath as not_unescaped;
+               ^
+select E'\'null \\u0000 escape\''::jsonpath as not_an_escape;
+ERROR:  unsupported Unicode escape sequence
+LINE 1: select E'\'null \\u0000 escape\''::jsonpath as not_an_escape...
+               ^
+DETAIL:  \u0000 cannot be converted to text.
+-- checks for quoted key names
+-- basic unicode input
+SELECT '$."\u"'::jsonpath;		-- ERROR, incomplete escape
+ERROR:  bad jsonpath representation
+LINE 1: SELECT '$."\u"'::jsonpath;
+               ^
+DETAIL:  Unicode sequence is invalid at or near "\u"
+SELECT '$."\u00"'::jsonpath;	-- ERROR, incomplete escape
+ERROR:  bad jsonpath representation
+LINE 1: SELECT '$."\u00"'::jsonpath;
+               ^
+DETAIL:  Unicode sequence is invalid at or near "\u00"
+SELECT '$."\u000g"'::jsonpath;	-- ERROR, g is not a hex digit
+ERROR:  bad jsonpath representation
+LINE 1: SELECT '$."\u000g"'::jsonpath;
+               ^
+DETAIL:  Unicode sequence is invalid at or near "\u000"
+SELECT '$."\u0000"'::jsonpath;	-- OK, legal escape
+ERROR:  unsupported Unicode escape sequence
+LINE 1: SELECT '$."\u0000"'::jsonpath;
+               ^
+DETAIL:  \u0000 cannot be converted to text.
+SELECT '$."\uaBcD"'::jsonpath;	-- OK, uppercase and lower case both OK
+ERROR:  invalid input syntax for type jsonpath
+LINE 1: SELECT '$."\uaBcD"'::jsonpath;
+               ^
+DETAIL:  Unicode escape values cannot be used for code point values above 007F when the server encoding is not UTF8.
+-- handling of unicode surrogate pairs
+select '$."\ud83d\ude04\ud83d\udc36"'::jsonpath as correct_in_utf8;
+ERROR:  invalid input syntax for type jsonpath
+LINE 1: select '$."\ud83d\ude04\ud83d\udc36"'::jsonpath as correct_i...
+               ^
+DETAIL:  Unicode escape values cannot be used for code point values above 007F when the server encoding is not UTF8.
+select '$."\ud83d\ud83d"'::jsonpath; -- 2 high surrogates in a row
+ERROR:  invalid input syntax for type jsonpath
+LINE 1: select '$."\ud83d\ud83d"'::jsonpath;
+               ^
+DETAIL:  Unicode high surrogate must not follow a high surrogate.
+select '$."\ude04\ud83d"'::jsonpath; -- surrogates in wrong order
+ERROR:  invalid input syntax for type jsonpath
+LINE 1: select '$."\ude04\ud83d"'::jsonpath;
+               ^
+DETAIL:  Unicode low surrogate must follow a high surrogate.
+select '$."\ud83dX"'::jsonpath; -- orphan high surrogate
+ERROR:  invalid input syntax for type jsonpath
+LINE 1: select '$."\ud83dX"'::jsonpath;
+               ^
+DETAIL:  Unicode low surrogate must follow a high surrogate.
+select '$."\ude04X"'::jsonpath; -- orphan low surrogate
+ERROR:  invalid input syntax for type jsonpath
+LINE 1: select '$."\ude04X"'::jsonpath;
+               ^
+DETAIL:  Unicode low surrogate must follow a high surrogate.
+--handling of simple unicode escapes
+select '$."the Copyright \u00a9 sign"'::jsonpath as correct_in_utf8;
+ERROR:  invalid input syntax for type jsonpath
+LINE 1: select '$."the Copyright \u00a9 sign"'::jsonpath as correct_...
+               ^
+DETAIL:  Unicode escape values cannot be used for code point values above 007F when the server encoding is not UTF8.
+select '$."dollar \u0024 character"'::jsonpath as correct_everywhere;
+   correct_everywhere   
+------------------------
+ $."dollar $ character"
+(1 row)
+
+select '$."dollar \\u0024 character"'::jsonpath as not_an_escape;
+        not_an_escape         
+------------------------------
+ $."dollar \\u0024 character"
+(1 row)
+
+select '$."null \u0000 escape"'::jsonpath as not_unescaped;
+ERROR:  unsupported Unicode escape sequence
+LINE 1: select '$."null \u0000 escape"'::jsonpath as not_unescaped;
+               ^
+DETAIL:  \u0000 cannot be converted to text.
+select '$."null \\u0000 escape"'::jsonpath as not_an_escape;
+      not_an_escape      
+-------------------------
+ $."null \\u0000 escape"
+(1 row)
+

src/test/regress/parallel_schedule

@@ -109,7 +109,7 @@ test: select_views portals_p2 foreign_key cluster dependency guc bitmapops combo
 # ----------
 # Another group of parallel tests (JSON related)
 # ----------
-test: json jsonb json_encoding jsonpath jsonb_jsonpath
+test: json jsonb json_encoding jsonpath jsonpath_encoding jsonb_jsonpath
 
 # ----------
 # Another group of parallel tests

src/test/regress/serial_schedule

@@ -160,6 +160,7 @@ test: json
 test: jsonb
 test: json_encoding
 test: jsonpath
+test: jsonpath_encoding
 test: jsonb_jsonpath
 test: indirect_toast
 test: equivclass

src/test/regress/sql/jsonb_jsonpath.sql

@@ -269,7 +269,7 @@ select jsonb_path_query('[null,1,true,"a",[],{}]', 'lax $.type()');
 select jsonb_path_query('[null,1,true,"a",[],{}]', '$[*].type()');
 select jsonb_path_query('null', 'null.type()');
 select jsonb_path_query('null', 'true.type()');
-select jsonb_path_query('null', '123.type()');
+select jsonb_path_query('null', '(123).type()');
 select jsonb_path_query('null', '"123".type()');
 
 select jsonb_path_query('{"a": 2}', '($.a - 5).abs() + 10');

src/test/regress/sql/jsonpath.sql

@@ -66,6 +66,8 @@ select '$[$[0] ? (last > 0)]'::jsonpath;
 
 select 'null.type()'::jsonpath;
 select '1.type()'::jsonpath;
+select '(1).type()'::jsonpath;
+select '1.2.type()'::jsonpath;
 select '"aaa".type()'::jsonpath;
 select 'true.type()'::jsonpath;
 select '$.double().floor().ceiling().abs()'::jsonpath;
@@ -145,3 +147,31 @@ select '$ ? (@.a < +0.1e+1)'::jsonpath;
 select '$ ? (@.a < 10.1e+1)'::jsonpath;
 select '$ ? (@.a < -10.1e+1)'::jsonpath;
 select '$ ? (@.a < +10.1e+1)'::jsonpath;
+
+select '0'::jsonpath;
+select '00'::jsonpath;
+select '0.0'::jsonpath;
+select '0.000'::jsonpath;
+select '0.000e1'::jsonpath;
+select '0.000e2'::jsonpath;
+select '0.000e3'::jsonpath;
+select '0.0010'::jsonpath;
+select '0.0010e-1'::jsonpath;
+select '0.0010e+1'::jsonpath;
+select '0.0010e+2'::jsonpath;
+select '1e'::jsonpath;
+select '1.e'::jsonpath;
+select '1.2e'::jsonpath;
+select '1.2.e'::jsonpath;
+select '(1.2).e'::jsonpath;
+select '1e3'::jsonpath;
+select '1.e3'::jsonpath;
+select '1.e3.e'::jsonpath;
+select '1.e3.e4'::jsonpath;
+select '1.2e3'::jsonpath;
+select '1.2.e3'::jsonpath;
+select '(1.2).e3'::jsonpath;
+select '1..e'::jsonpath;
+select '1..e3'::jsonpath;
+select '(1.).e'::jsonpath;
+select '(1.).e3'::jsonpath;

src/test/regress/sql/jsonpath_encoding.sql

@@ -0,0 +1,71 @@
+
+-- encoding-sensitive tests for jsonpath
+
+-- checks for double-quoted values
+
+-- basic unicode input
+SELECT '"\u"'::jsonpath;		-- ERROR, incomplete escape
+SELECT '"\u00"'::jsonpath;		-- ERROR, incomplete escape
+SELECT '"\u000g"'::jsonpath;	-- ERROR, g is not a hex digit
+SELECT '"\u0000"'::jsonpath;	-- OK, legal escape
+SELECT '"\uaBcD"'::jsonpath;	-- OK, uppercase and lower case both OK
+
+-- handling of unicode surrogate pairs
+select '"\ud83d\ude04\ud83d\udc36"'::jsonpath as correct_in_utf8;
+select '"\ud83d\ud83d"'::jsonpath; -- 2 high surrogates in a row
+select '"\ude04\ud83d"'::jsonpath; -- surrogates in wrong order
+select '"\ud83dX"'::jsonpath; -- orphan high surrogate
+select '"\ude04X"'::jsonpath; -- orphan low surrogate
+
+--handling of simple unicode escapes
+select '"the Copyright \u00a9 sign"'::jsonpath as correct_in_utf8;
+select '"dollar \u0024 character"'::jsonpath as correct_everywhere;
+select '"dollar \\u0024 character"'::jsonpath as not_an_escape;
+select '"null \u0000 escape"'::jsonpath as not_unescaped;
+select '"null \\u0000 escape"'::jsonpath as not_an_escape;
+
+-- checks for single-quoted values
+
+-- basic unicode input
+SELECT E'\'\u\''::jsonpath;		-- ERROR, incomplete escape
+SELECT E'\'\u00\''::jsonpath;	-- ERROR, incomplete escape
+SELECT E'\'\u000g\''::jsonpath;	-- ERROR, g is not a hex digit
+SELECT E'\'\u0000\''::jsonpath;	-- OK, legal escape
+SELECT E'\'\uaBcD\''::jsonpath;	-- OK, uppercase and lower case both OK
+
+-- handling of unicode surrogate pairs
+select E'\'\ud83d\ude04\ud83d\udc36\''::jsonpath as correct_in_utf8;
+select E'\'\ud83d\ud83d\''::jsonpath; -- 2 high surrogates in a row
+select E'\'\ude04\ud83d\''::jsonpath; -- surrogates in wrong order
+select E'\'\ud83dX\''::jsonpath; -- orphan high surrogate
+select E'\'\ude04X\''::jsonpath; -- orphan low surrogate
+
+--handling of simple unicode escapes
+select E'\'the Copyright \u00a9 sign\''::jsonpath as correct_in_utf8;
+select E'\'dollar \u0024 character\''::jsonpath as correct_everywhere;
+select E'\'dollar \\u0024 character\''::jsonpath as not_an_escape;
+select E'\'null \u0000 escape\''::jsonpath as not_unescaped;
+select E'\'null \\u0000 escape\''::jsonpath as not_an_escape;
+
+-- checks for quoted key names
+
+-- basic unicode input
+SELECT '$."\u"'::jsonpath;		-- ERROR, incomplete escape
+SELECT '$."\u00"'::jsonpath;	-- ERROR, incomplete escape
+SELECT '$."\u000g"'::jsonpath;	-- ERROR, g is not a hex digit
+SELECT '$."\u0000"'::jsonpath;	-- OK, legal escape
+SELECT '$."\uaBcD"'::jsonpath;	-- OK, uppercase and lower case both OK
+
+-- handling of unicode surrogate pairs
+select '$."\ud83d\ude04\ud83d\udc36"'::jsonpath as correct_in_utf8;
+select '$."\ud83d\ud83d"'::jsonpath; -- 2 high surrogates in a row
+select '$."\ude04\ud83d"'::jsonpath; -- surrogates in wrong order
+select '$."\ud83dX"'::jsonpath; -- orphan high surrogate
+select '$."\ude04X"'::jsonpath; -- orphan low surrogate
+
+--handling of simple unicode escapes
+select '$."the Copyright \u00a9 sign"'::jsonpath as correct_in_utf8;
+select '$."dollar \u0024 character"'::jsonpath as correct_everywhere;
+select '$."dollar \\u0024 character"'::jsonpath as not_an_escape;
+select '$."null \u0000 escape"'::jsonpath as not_unescaped;
+select '$."null \\u0000 escape"'::jsonpath as not_an_escape;