From 562f29aedb0207420a8e3aeb807a9a1b18095fda Mon Sep 17 00:00:00 2001 From: Tom Lane Date: Tue, 30 Dec 2025 15:42:22 -0500 Subject: [PATCH v1 2/4] Remove server-side support for !standard_conforming_strings. Mostly this consists of dropping scan.l's support for escape_string_warning, which is no longer reachable. (To see that, note that the warning subroutines are only called within the exclusive state, and after the change in the {xqstart} rule, state cannot be reached with yyextra->warn_on_first_escape true. So the subroutines can never warn.) This leaves the escape_string_warning GUC nonfunctional, but cleaning up that situation will come later. Also adjust assorted comments to reflect the fact that standard_conforming_strings = false is no longer possible locally, but we still have to allow for the possibility that a remote server might have it. Author: Tom Lane Discussion: https://postgr.es/m/3279216.1767072538@sss.pgh.pa.us --- .../pg_stat_statements/pg_stat_statements.c | 3 - contrib/test_decoding/test_decoding.c | 4 +- src/backend/parser/scan.l | 76 ++----------------- src/backend/utils/adt/quote.c | 8 +- src/backend/utils/adt/ruleutils.c | 6 +- src/backend/utils/misc/guc_tables.c | 5 +- src/include/parser/parser.h | 4 +- src/include/parser/scanner.h | 5 +- 8 files changed, 17 insertions(+), 94 deletions(-) diff --git a/contrib/pg_stat_statements/pg_stat_statements.c b/contrib/pg_stat_statements/pg_stat_statements.c index 39208f80b5b..58892d12246 100644 --- a/contrib/pg_stat_statements/pg_stat_statements.c +++ b/contrib/pg_stat_statements/pg_stat_statements.c @@ -2991,9 +2991,6 @@ fill_in_constant_lengths(JumbleState *jstate, const char *query, &ScanKeywords, ScanKeywordTokens); - /* we don't want to re-emit any escape string warnings */ - yyextra.escape_string_warning = false; - /* Search for each constant, in sequence */ for (int i = 0; i < jstate->clocations_count; i++) { diff --git a/contrib/test_decoding/test_decoding.c b/contrib/test_decoding/test_decoding.c index 47094f86f5f..4c92f0eeffa 100644 --- a/contrib/test_decoding/test_decoding.c +++ b/contrib/test_decoding/test_decoding.c @@ -474,8 +474,8 @@ pg_decode_filter(LogicalDecodingContext *ctx, * Print literal `outputstr' already represented as string of type `typid' * into stringbuf `s'. * - * Some builtin types aren't quoted, the rest is quoted. Escaping is done as - * if standard_conforming_strings were enabled. + * Some builtin types aren't quoted, the rest is quoted. Escaping is done + * per standard SQL rules. */ static void print_literal(StringInfo s, Oid typid, char *outputstr) diff --git a/src/backend/parser/scan.l b/src/backend/parser/scan.l index a67815339b7..cc88eea3708 100644 --- a/src/backend/parser/scan.l +++ b/src/backend/parser/scan.l @@ -60,14 +60,13 @@ fprintf_to_ereport(const char *fmt, const char *msg) } /* - * GUC variables. This is a DIRECT violation of the warning given at the + * GUC variable. This is a DIRECT violation of the warning given at the * head of gram.y, ie flex/bison code must not depend on any GUC variables; - * as such, changing their values can induce very unintuitive behavior. - * But we shall have to live with it until we can remove these variables. + * as such, changing its value can induce very unintuitive behavior. + * In practice, backslash_quote is not too awful since it only controls + * whether to throw an error: it cannot change non-error results. */ int backslash_quote = BACKSLASH_QUOTE_SAFE_ENCODING; -bool escape_string_warning = true; -bool standard_conforming_strings = true; /* * Constant data exported from this file. This array maps from the @@ -127,9 +126,6 @@ static void addunicode(char32_t c, yyscan_t yyscanner); #define lexer_errposition() scanner_errposition(*(yylloc), yyscanner) -static void check_string_escape_warning(unsigned char ychar, core_yyscan_t yyscanner); -static void check_escape_warning(core_yyscan_t yyscanner); - %} %option reentrant @@ -543,17 +539,12 @@ other . } {xqstart} { - yyextra->warn_on_first_escape = true; yyextra->saw_non_ascii = false; SET_YYLLOC(); - if (yyextra->standard_conforming_strings) - BEGIN(xq); - else - BEGIN(xe); + BEGIN(xq); startlit(); } {xestart} { - yyextra->warn_on_first_escape = false; yyextra->saw_non_ascii = false; SET_YYLLOC(); BEGIN(xe); @@ -561,12 +552,6 @@ other . } {xusstart} { SET_YYLLOC(); - if (!yyextra->standard_conforming_strings) - ereport(ERROR, - (errcode(ERRCODE_FEATURE_NOT_SUPPORTED), - errmsg("unsafe use of string constant with Unicode escapes"), - errdetail("String constants with Unicode escapes cannot be used when \"standard_conforming_strings\" is off."), - lexer_errposition())); BEGIN(xus); startlit(); } @@ -642,13 +627,6 @@ other . {xeunicode} { char32_t c = strtoul(yytext + 2, NULL, 16); - /* - * For consistency with other productions, issue any - * escape warning with cursor pointing to start of string. - * We might want to change that, someday. - */ - check_escape_warning(yyscanner); - /* Remember start of overall string token ... */ PUSH_YYLLOC(); /* ... and set the error cursor to point at this esc seq */ @@ -715,14 +693,12 @@ other . errhint("Use '' to write quotes in strings. \\' is insecure in client-only encodings."), lexer_errposition())); } - check_string_escape_warning(yytext[1], yyscanner); addlitchar(unescape_single_char(yytext[1], yyscanner), yyscanner); } {xeoctesc} { unsigned char c = strtoul(yytext + 1, NULL, 8); - check_escape_warning(yyscanner); addlitchar(c, yyscanner); if (c == '\0' || IS_HIGHBIT_SET(c)) yyextra->saw_non_ascii = true; @@ -730,7 +706,6 @@ other . {xehexesc} { unsigned char c = strtoul(yytext + 2, NULL, 16); - check_escape_warning(yyscanner); addlitchar(c, yyscanner); if (c == '\0' || IS_HIGHBIT_SET(c)) yyextra->saw_non_ascii = true; @@ -1263,8 +1238,6 @@ scanner_init(const char *str, yyext->keyword_tokens = keyword_tokens; yyext->backslash_quote = backslash_quote; - yyext->escape_string_warning = escape_string_warning; - yyext->standard_conforming_strings = standard_conforming_strings; /* * Make a scan buffer with special termination needed by flex. @@ -1420,45 +1393,6 @@ unescape_single_char(unsigned char c, core_yyscan_t yyscanner) } } -static void -check_string_escape_warning(unsigned char ychar, core_yyscan_t yyscanner) -{ - if (ychar == '\'') - { - if (yyextra->warn_on_first_escape && yyextra->escape_string_warning) - ereport(WARNING, - (errcode(ERRCODE_NONSTANDARD_USE_OF_ESCAPE_CHARACTER), - errmsg("nonstandard use of \\' in a string literal"), - errhint("Use '' to write quotes in strings, or use the escape string syntax (E'...')."), - lexer_errposition())); - yyextra->warn_on_first_escape = false; /* warn only once per string */ - } - else if (ychar == '\\') - { - if (yyextra->warn_on_first_escape && yyextra->escape_string_warning) - ereport(WARNING, - (errcode(ERRCODE_NONSTANDARD_USE_OF_ESCAPE_CHARACTER), - errmsg("nonstandard use of \\\\ in a string literal"), - errhint("Use the escape string syntax for backslashes, e.g., E'\\\\'."), - lexer_errposition())); - yyextra->warn_on_first_escape = false; /* warn only once per string */ - } - else - check_escape_warning(yyscanner); -} - -static void -check_escape_warning(core_yyscan_t yyscanner) -{ - if (yyextra->warn_on_first_escape && yyextra->escape_string_warning) - ereport(WARNING, - (errcode(ERRCODE_NONSTANDARD_USE_OF_ESCAPE_CHARACTER), - errmsg("nonstandard use of escape in a string literal"), - errhint("Use the escape string syntax for escapes, e.g., E'\\r\\n'."), - lexer_errposition())); - yyextra->warn_on_first_escape = false; /* warn only once per string */ -} - /* * Interface functions to make flex use palloc() instead of malloc(). * It'd be better to make these static, but flex insists otherwise. diff --git a/src/backend/utils/adt/quote.c b/src/backend/utils/adt/quote.c index 551de59a07f..e62bf04a9f9 100644 --- a/src/backend/utils/adt/quote.c +++ b/src/backend/utils/adt/quote.c @@ -37,11 +37,9 @@ quote_ident(PG_FUNCTION_ARGS) * quote_literal_internal - * helper function for quote_literal and quote_literal_cstr * - * NOTE: think not to make this function's behavior change with - * standard_conforming_strings. We don't know where the result - * literal will be used, and so we must generate a result that - * will work with either setting. Take a look at what dblink - * uses this for before thinking you know better. + * NOTE: This must produce output that will work in old servers with + * standard_conforming_strings = off. It's used for example by + * dblink, which may send the result to another server. */ static size_t quote_literal_internal(char *dst, const char *src, size_t len) diff --git a/src/backend/utils/adt/ruleutils.c b/src/backend/utils/adt/ruleutils.c index 9f85eb86da1..61028db020f 100644 --- a/src/backend/utils/adt/ruleutils.c +++ b/src/backend/utils/adt/ruleutils.c @@ -11828,16 +11828,14 @@ simple_quote_literal(StringInfo buf, const char *val) const char *valptr; /* - * We form the string literal according to the prevailing setting of - * standard_conforming_strings; we never use E''. User is responsible for - * making sure result is used correctly. + * We always form the string literal according to standard SQL rules. */ appendStringInfoChar(buf, '\''); for (valptr = val; *valptr; valptr++) { char ch = *valptr; - if (SQL_STR_DOUBLE(ch, !standard_conforming_strings)) + if (SQL_STR_DOUBLE(ch, false)) appendStringInfoChar(buf, ch); appendStringInfoChar(buf, ch); } diff --git a/src/backend/utils/misc/guc_tables.c b/src/backend/utils/misc/guc_tables.c index 04ab0a26608..beea9b00657 100644 --- a/src/backend/utils/misc/guc_tables.c +++ b/src/backend/utils/misc/guc_tables.c @@ -529,10 +529,11 @@ bool row_security; bool check_function_bodies = true; /* - * This GUC exists solely for backward compatibility, check its definition for - * details. + * These GUCs exist solely for backward compatibility. */ static bool default_with_oids = false; +static bool standard_conforming_strings = true; +static bool escape_string_warning = true; bool current_role_is_superuser; diff --git a/src/include/parser/parser.h b/src/include/parser/parser.h index 350196cd641..cbf9a704dfb 100644 --- a/src/include/parser/parser.h +++ b/src/include/parser/parser.h @@ -52,10 +52,8 @@ typedef enum BACKSLASH_QUOTE_SAFE_ENCODING, } BackslashQuoteType; -/* GUC variables in scan.l (every one of these is a bad idea :-() */ +/* GUC variable in scan.l */ extern PGDLLIMPORT int backslash_quote; -extern PGDLLIMPORT bool escape_string_warning; -extern PGDLLIMPORT bool standard_conforming_strings; /* Primary entry point for the raw parsing functions */ diff --git a/src/include/parser/scanner.h b/src/include/parser/scanner.h index 8d202d5b284..f5806feb91d 100644 --- a/src/include/parser/scanner.h +++ b/src/include/parser/scanner.h @@ -85,8 +85,6 @@ typedef struct core_yy_extra_type * prevailing GUC settings. */ int backslash_quote; - bool escape_string_warning; - bool standard_conforming_strings; /* * literalbuf is used to accumulate literal values when multiple rules are @@ -110,8 +108,7 @@ typedef struct core_yy_extra_type /* first part of UTF16 surrogate pair for Unicode escapes */ int32 utf16_first_part; - /* state variables for literal-lexing warnings */ - bool warn_on_first_escape; + /* true if we need to verify valid encoding of current literal string */ bool saw_non_ascii; } core_yy_extra_type; -- 2.43.7