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

3 files changed, 102 insertions, 1 deletions

diff --git a/src/test/regress/sql/jsonb_jsonpath.sql b/src/test/regress/sql/jsonb_jsonpath.sql
index 41b346b2d4d..28c861bb179 100644
--- a/src/test/regress/sql/jsonb_jsonpath.sql
+++ b/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}', '($.a - 5).abs() + 10');
diff --git a/src/test/regress/sql/jsonpath.sql b/src/test/regress/sql/jsonpath.sql
index e5f3391a666..9171ddbc6cd 100644
--- a/src/test/regress/sql/jsonpath.sql
+++ b/src/test/regress/sql/jsonpath.sql
@@ -66,6 +66,8 @@ select '$[$[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;
diff --git a/src/test/regress/sql/jsonpath_encoding.sql b/src/test/regress/sql/jsonpath_encoding.sql
new file mode 100644
index 00000000000..a3b5bc39a1c
--- /dev/null
+++ b/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;