Skip to content

Commit 2549f06

Browse files
peterepetere
committed
Reject trailing junk after numeric literals
After this, the PostgreSQL lexers no longer accept numeric literals with trailing non-digits, such as 123abc, which would be scanned as two tokens: 123 and abc. This is undocumented and surprising, and it might also interfere with some extended numeric literal syntax being contemplated for the future. Reviewed-by: John Naylor <john.naylor@enterprisedb.com> Discussion: https://www.postgresql.org/message-id/flat/b239564c-cad0-b23e-c57e-166d883cb97d@enterprisedb.com
1 parent 70e8186 commit 2549f06

File tree

5 files changed

+96
-86
lines changed

5 files changed

+96
-86
lines changed

‎src/backend/parser/scan.l‎

Lines changed: 23 additions & 14 deletions
Original file line numberDiff line numberDiff line change
@@ -387,7 +387,7 @@ operator {op_chars}+
387387
*
388388
* {decimalfail} is used because we would like "1..10" to lex as 1, dot_dot, 10.
389389
*
390-
* {realfail1} and {realfail2} are added to prevent the need for scanner
390+
* {realfail} is added to prevent the need for scanner
391391
* backup when the {real} rule fails to match completely.
392392
*/
393393
digit [0-9]
@@ -396,10 +396,14 @@ integer {digit}+
396396
decimal (({digit}*\.{digit}+)|({digit}+\.{digit}*))
397397
decimalfail {digit}+\.\.
398398
real ({integer}|{decimal})[Ee][-+]?{digit}+
399-
realfail1 ({integer}|{decimal})[Ee]
400-
realfail2 ({integer}|{decimal})[Ee][-+]
399+
realfail ({integer}|{decimal})[Ee][-+]
400+
401+
integer_junk {integer}{ident_start}
402+
decimal_junk {decimal}{ident_start}
403+
real_junk {real}{ident_start}
401404

402405
param \${integer}
406+
param_junk \${integer}{ident_start}
403407

404408
other .
405409

@@ -974,6 +978,10 @@ other .
974978
yylval->ival = atol(yytext + 1);
975979
return PARAM;
976980
}
981+
{param_junk} {
982+
SET_YYLLOC();
983+
yyerror("trailing junk after parameter");
984+
}
977985

978986
{integer} {
979987
SET_YYLLOC();
@@ -995,20 +1003,21 @@ other .
9951003
yylval->str = pstrdup(yytext);
9961004
return FCONST;
9971005
}
998-
{realfail1} {
999-
/*
1000-
* throw back the [Ee], and figure out whether what
1001-
* remains is an {integer} or {decimal}.
1002-
*/
1003-
yyless(yyleng - 1);
1006+
{realfail} {
10041007
SET_YYLLOC();
1005-
return process_integer_literal(yytext, yylval);
1008+
yyerror("trailing junk after numeric literal");
10061009
}
1007-
{realfail2} {
1008-
/* throw back the [Ee][+-], and proceed as above */
1009-
yyless(yyleng - 2);
1010+
{integer_junk} {
10101011
SET_YYLLOC();
1011-
return process_integer_literal(yytext, yylval);
1012+
yyerror("trailing junk after numeric literal");
1013+
}
1014+
{decimal_junk} {
1015+
SET_YYLLOC();
1016+
yyerror("trailing junk after numeric literal");
1017+
}
1018+
{real_junk} {
1019+
SET_YYLLOC();
1020+
yyerror("trailing junk after numeric literal");
10121021
}
10131022

10141023

‎src/fe_utils/psqlscan.l‎

Lines changed: 18 additions & 13 deletions
Original file line numberDiff line numberDiff line change
@@ -325,7 +325,7 @@ operator {op_chars}+
325325
*
326326
* {decimalfail} is used because we would like "1..10" to lex as 1, dot_dot, 10.
327327
*
328-
* {realfail1} and {realfail2} are added to prevent the need for scanner
328+
* {realfail} is added to prevent the need for scanner
329329
* backup when the {real} rule fails to match completely.
330330
*/
331331
digit [0-9]
@@ -334,10 +334,14 @@ integer {digit}+
334334
decimal (({digit}*\.{digit}+)|({digit}+\.{digit}*))
335335
decimalfail {digit}+\.\.
336336
real ({integer}|{decimal})[Ee][-+]?{digit}+
337-
realfail1 ({integer}|{decimal})[Ee]
338-
realfail2 ({integer}|{decimal})[Ee][-+]
337+
realfail ({integer}|{decimal})[Ee][-+]
338+
339+
integer_junk {integer}{ident_start}
340+
decimal_junk {decimal}{ident_start}
341+
real_junk {real}{ident_start}
339342

340343
param \${integer}
344+
param_junk \${integer}{ident_start}
341345

342346
/* psql-specific: characters allowed in variable names */
343347
variable_char [A-Za-z\200-\377_0-9]
@@ -839,6 +843,9 @@ other .
839843
{param} {
840844
ECHO;
841845
}
846+
{param_junk} {
847+
ECHO;
848+
}
842849

843850
{integer} {
844851
ECHO;
@@ -854,18 +861,16 @@ other .
854861
{real} {
855862
ECHO;
856863
}
857-
{realfail1} {
858-
/*
859-
* throw back the [Ee], and figure out whether what
860-
* remains is an {integer} or {decimal}.
861-
* (in psql, we don't actually care...)
862-
*/
863-
yyless(yyleng - 1);
864+
{realfail} {
864865
ECHO;
865866
}
866-
{realfail2} {
867-
/* throw back the [Ee][+-], and proceed as above */
868-
yyless(yyleng - 2);
867+
{integer_junk} {
868+
ECHO;
869+
}
870+
{decimal_junk} {
871+
ECHO;
872+
}
873+
{real_junk} {
869874
ECHO;
870875
}
871876

‎src/interfaces/ecpg/preproc/pgc.l‎

Lines changed: 26 additions & 10 deletions
Original file line numberDiff line numberDiff line change
@@ -353,7 +353,7 @@ operator {op_chars}+
353353
*
354354
* {decimalfail} is used because we would like "1..10" to lex as 1, dot_dot, 10.
355355
*
356-
* {realfail1} and {realfail2} are added to prevent the need for scanner
356+
* {realfail} is added to prevent the need for scanner
357357
* backup when the {real} rule fails to match completely.
358358
*/
359359
digit [0-9]
@@ -362,10 +362,14 @@ integer {digit}+
362362
decimal (({digit}*\.{digit}+)|({digit}+\.{digit}*))
363363
decimalfail {digit}+\.\.
364364
real ({integer}|{decimal})[Ee][-+]?{digit}+
365-
realfail1 ({integer}|{decimal})[Ee]
366-
realfail2 ({integer}|{decimal})[Ee][-+]
365+
realfail ({integer}|{decimal})[Ee][-+]
366+
367+
integer_junk {integer}{ident_start}
368+
decimal_junk {decimal}{ident_start}
369+
real_junk {real}{ident_start}
367370

368371
param \${integer}
372+
param_junk \${integer}{ident_start}
369373

370374
/* special characters for other dbms */
371375
/* we have to react differently in compat mode */
@@ -917,6 +921,9 @@ cppline {space}*#([^i][A-Za-z]*|{if}|{ifdef}|{ifndef}|{import})((\/\*[^*/]*\*+
917921
base_yylval.ival = atol(yytext+1);
918922
return PARAM;
919923
}
924+
{param_junk} {
925+
mmfatal(PARSE_ERROR, "trailing junk after parameter");
926+
}
920927

921928
{ip} {
922929
base_yylval.str = mm_strdup(yytext);
@@ -941,22 +948,31 @@ cppline {space}*#([^i][A-Za-z]*|{if}|{ifdef}|{ifndef}|{import})((\/\*[^*/]*\*+
941948
base_yylval.str = mm_strdup(yytext);
942949
return FCONST;
943950
}
944-
{realfail1} {
951+
{realfail} {
945952
/*
946-
* throw back the [Ee], and figure out whether what
953+
* throw back the [Ee][+-], and figure out whether what
947954
* remains is an {integer} or {decimal}.
948955
*/
949-
yyless(yyleng - 1);
950-
return process_integer_literal(yytext, &base_yylval);
951-
}
952-
{realfail2} {
953-
/* throw back the [Ee][+-], and proceed as above */
954956
yyless(yyleng - 2);
955957
return process_integer_literal(yytext, &base_yylval);
956958
}
957959
} /* <C,SQL> */
958960

959961
<SQL>{
962+
/*
963+
* Note that some trailing junk is valid in C (such as 100LL), so we
964+
* contain this to SQL mode.
965+
*/
966+
{integer_junk} {
967+
mmfatal(PARSE_ERROR, "trailing junk after numeric literal");
968+
}
969+
{decimal_junk} {
970+
mmfatal(PARSE_ERROR, "trailing junk after numeric literal");
971+
}
972+
{real_junk} {
973+
mmfatal(PARSE_ERROR, "trailing junk after numeric literal");
974+
}
975+
960976
:{identifier}((("->"|\.){identifier})|(\[{array}\]))* {
961977
base_yylval.str = mm_strdup(yytext+1);
962978
return CVARIABLE;

‎src/test/regress/expected/numerology.out‎

Lines changed: 29 additions & 48 deletions
Original file line numberDiff line numberDiff line change
@@ -6,64 +6,45 @@
66
-- Trailing junk in numeric literals
77
--
88
SELECT 123abc;
9-
abc
10-
-----
11-
123
12-
(1 row)
13-
9+
ERROR: trailing junk after numeric literal at or near "123a"
10+
LINE 1: SELECT 123abc;
11+
^
1412
SELECT 0x0o;
15-
x0o
16-
-----
17-
0
18-
(1 row)
19-
13+
ERROR: trailing junk after numeric literal at or near "0x"
14+
LINE 1: SELECT 0x0o;
15+
^
2016
SELECT 1_2_3;
21-
_2_3
22-
------
23-
1
24-
(1 row)
25-
17+
ERROR: trailing junk after numeric literal at or near "1_"
18+
LINE 1: SELECT 1_2_3;
19+
^
2620
SELECT 0.a;
27-
a
28-
---
29-
0
30-
(1 row)
31-
21+
ERROR: trailing junk after numeric literal at or near "0.a"
22+
LINE 1: SELECT 0.a;
23+
^
3224
SELECT 0.0a;
33-
a
34-
-----
35-
0.0
36-
(1 row)
37-
25+
ERROR: trailing junk after numeric literal at or near "0.0a"
26+
LINE 1: SELECT 0.0a;
27+
^
3828
SELECT .0a;
39-
a
40-
-----
41-
0.0
42-
(1 row)
43-
29+
ERROR: trailing junk after numeric literal at or near ".0a"
30+
LINE 1: SELECT .0a;
31+
^
4432
SELECT 0.0e1a;
45-
a
46-
---
47-
0
48-
(1 row)
49-
33+
ERROR: trailing junk after numeric literal at or near "0.0e1a"
34+
LINE 1: SELECT 0.0e1a;
35+
^
5036
SELECT 0.0e;
51-
e
52-
-----
53-
0.0
54-
(1 row)
55-
37+
ERROR: trailing junk after numeric literal at or near "0.0e"
38+
LINE 1: SELECT 0.0e;
39+
^
5640
SELECT 0.0e+a;
57-
ERROR: syntax error at or near "+"
41+
ERROR: trailing junk after numeric literal at or near "0.0e+"
5842
LINE 1: SELECT 0.0e+a;
59-
^
43+
^
6044
PREPARE p1 AS SELECT $1a;
61-
EXECUTE p1(1);
62-
a
63-
---
64-
1
65-
(1 row)
66-
45+
ERROR: trailing junk after parameter at or near "$1a"
46+
LINE 1: PREPARE p1 AS SELECT $1a;
47+
^
6748
--
6849
-- Test implicit type conversions
6950
-- This fails for Postgres v6.1 (and earlier?)

‎src/test/regress/sql/numerology.sql‎

Lines changed: 0 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -17,7 +17,6 @@ SELECT 0.0e1a;
1717
SELECT 0.0e;
1818
SELECT 0.0e+a;
1919
PREPARE p1 AS SELECT $1a;
20-
EXECUTE p1(1);
2120

2221
--
2322
-- Test implicit type conversions

0 commit comments

Comments
 (0)