bootstrap: pure parser and reentrant scanner

Use the flex %option reentrant and the bison option %pure-parser to
make the generated scanner and parser pure, reentrant, and
thread-safe.

Make the generated scanner use palloc() etc. instead of malloc() etc.

For the bootstrap scanner and parser, reentrancy and memory management
aren't that important, but we make this change here anyway so that all
the scanners and parsers in the backend use a similar set of options
and APIs.

Reviewed-by: Heikki Linnakangas <hlinnaka@iki.fi>
Reviewed-by: Andreas Karlsson <andreas@proxel.se>
Discussion: https://www.postgresql.org/message-id/flat/eb6faeac-2a8a-4b69-9189-c33c520e5b7b@eisentraut.org

src/backend/bootstrap/bootparse.y

@@ -34,10 +34,6 @@
 
 #include "bootparse.h"
 
-/* silence -Wmissing-variable-declarations */
-extern int boot_yychar;
-extern int boot_yynerrs;
-
 
 /*
  * Bison doesn't allocate anything that needs to live across parser calls,
@@ -81,6 +77,9 @@ static int num_columns_read = 0;
 
 %}
 
+%parse-param {yyscan_t yyscanner}
+%lex-param   {yyscan_t yyscanner}
+%pure-parser
 %expect 0
 %name-prefix="boot_yy"
 
@@ -141,6 +140,8 @@ Boot_OpenStmt:
 					do_start();
 					boot_openrel($2);
 					do_end();
+
+					(void) yynerrs; /* suppress compiler warning */
 				}
 		;
 

src/backend/bootstrap/bootscanner.l

@@ -39,17 +39,19 @@ fprintf_to_ereport(const char *fmt, const char *msg)
 	ereport(ERROR, (errmsg_internal("%s", msg)));
 }
 
-
-static int	yyline = 1;			/* line number for error reporting */
-
 %}
 
+%option reentrant
+%option bison-bridge
 %option 8bit
 %option never-interactive
 %option nodefault
 %option noinput
 %option nounput
 %option noyywrap
+%option noyyalloc
+%option noyyrealloc
+%option noyyfree
 %option warn
 %option prefix="boot_yy"
 
@@ -58,7 +60,7 @@ id		[-A-Za-z0-9_]+
 sid		\'([^']|\'\')*\'
 
 /*
- * Keyword tokens return the keyword text (as a constant string) in boot_yylval.kw,
+ * Keyword tokens return the keyword text (as a constant string) in yylval->kw,
  * just in case that's needed because we want to treat the keyword as an
  * unreserved identifier.  Note that _null_ is not treated as a keyword
  * for this purpose; it's the one "reserved word" in the bootstrap syntax.
@@ -66,23 +68,23 @@ sid		\'([^']|\'\')*\'
  * Notice that all the keywords are case-sensitive, and for historical
  * reasons some must be upper case.
  *
- * String tokens return a palloc'd string in boot_yylval.str.
+ * String tokens return a palloc'd string in yylval->str.
  */
 
 %%
 
-open			{ boot_yylval.kw = "open"; return OPEN; }
+open			{ yylval->kw = "open"; return OPEN; }
 
-close			{ boot_yylval.kw = "close"; return XCLOSE; }
+close			{ yylval->kw = "close"; return XCLOSE; }
 
-create			{ boot_yylval.kw = "create"; return XCREATE; }
+create			{ yylval->kw = "create"; return XCREATE; }
 
-OID				{ boot_yylval.kw = "OID"; return OBJ_ID; }
-bootstrap		{ boot_yylval.kw = "bootstrap"; return XBOOTSTRAP; }
-shared_relation	{ boot_yylval.kw = "shared_relation"; return XSHARED_RELATION; }
-rowtype_oid		{ boot_yylval.kw = "rowtype_oid"; return XROWTYPE_OID; }
+OID				{ yylval->kw = "OID"; return OBJ_ID; }
+bootstrap		{ yylval->kw = "bootstrap"; return XBOOTSTRAP; }
+shared_relation	{ yylval->kw = "shared_relation"; return XSHARED_RELATION; }
+rowtype_oid		{ yylval->kw = "rowtype_oid"; return XROWTYPE_OID; }
 
-insert			{ boot_yylval.kw = "insert"; return INSERT_TUPLE; }
+insert			{ yylval->kw = "insert"; return INSERT_TUPLE; }
 
 _null_			{ return NULLVAL; }
 
@@ -91,35 +93,35 @@ _null_			{ return NULLVAL; }
 "("				{ return LPAREN; }
 ")"				{ return RPAREN; }
 
-[\n]			{ yyline++; }
+[\n]			{ yylineno++; }
 [\r\t ]			;
 
 ^\#[^\n]*		;		/* drop everything after "#" for comments */
 
-declare			{ boot_yylval.kw = "declare"; return XDECLARE; }
-build			{ boot_yylval.kw = "build"; return XBUILD; }
-indices			{ boot_yylval.kw = "indices"; return INDICES; }
-unique			{ boot_yylval.kw = "unique"; return UNIQUE; }
-index			{ boot_yylval.kw = "index"; return INDEX; }
-on				{ boot_yylval.kw = "on"; return ON; }
-using			{ boot_yylval.kw = "using"; return USING; }
-toast			{ boot_yylval.kw = "toast"; return XTOAST; }
-FORCE			{ boot_yylval.kw = "FORCE"; return XFORCE; }
-NOT				{ boot_yylval.kw = "NOT"; return XNOT; }
-NULL			{ boot_yylval.kw = "NULL"; return XNULL; }
+declare			{ yylval->kw = "declare"; return XDECLARE; }
+build			{ yylval->kw = "build"; return XBUILD; }
+indices			{ yylval->kw = "indices"; return INDICES; }
+unique			{ yylval->kw = "unique"; return UNIQUE; }
+index			{ yylval->kw = "index"; return INDEX; }
+on				{ yylval->kw = "on"; return ON; }
+using			{ yylval->kw = "using"; return USING; }
+toast			{ yylval->kw = "toast"; return XTOAST; }
+FORCE			{ yylval->kw = "FORCE"; return XFORCE; }
+NOT				{ yylval->kw = "NOT"; return XNOT; }
+NULL			{ yylval->kw = "NULL"; return XNULL; }
 
 {id}			{
-					boot_yylval.str = pstrdup(yytext);
+					yylval->str = pstrdup(yytext);
 					return ID;
 				}
 {sid}			{
 					/* strip quotes and escapes */
-					boot_yylval.str = DeescapeQuotedString(yytext);
+					yylval->str = DeescapeQuotedString(yytext);
 					return ID;
 				}
 
 .				{
-					elog(ERROR, "syntax error at line %d: unexpected character \"%s\"", yyline, yytext);
+					elog(ERROR, "syntax error at line %d: unexpected character \"%s\"", yylineno, yytext);
 				}
 
 %%
@@ -127,7 +129,36 @@ NULL			{ boot_yylval.kw = "NULL"; return XNULL; }
 /* LCOV_EXCL_STOP */
 
 void
-boot_yyerror(const char *message)
+boot_yyerror(yyscan_t yyscanner, const char *message)
+{
+	struct yyguts_t * yyg = (struct yyguts_t *) yyscanner;	/* needed for yylineno macro */
+
+	elog(ERROR, "%s at line %d", message, yylineno);
+}
+
+/*
+ * Interface functions to make flex use palloc() instead of malloc().
+ * It'd be better to make these static, but flex insists otherwise.
+ */
+
+void *
+yyalloc(yy_size_t size, yyscan_t yyscanner)
+{
+	return palloc(size);
+}
+
+void *
+yyrealloc(void *ptr, yy_size_t size, yyscan_t yyscanner)
+{
+	if (ptr)
+		return repalloc(ptr, size);
+	else
+		return palloc(size);
+}
+
+void
+yyfree(void *ptr, yyscan_t yyscanner)
 {
-	elog(ERROR, "%s at line %d", message, yyline);
+   if (ptr)
+	   pfree(ptr);
 }

src/backend/bootstrap/bootstrap.c

@@ -202,6 +202,7 @@ BootstrapModeMain(int argc, char *argv[], bool check_only)
 	int			flag;
 	char	   *userDoption = NULL;
 	uint32		bootstrap_data_checksum_version = 0;	/* No checksum */
+	yyscan_t	scanner;
 
 	Assert(!IsUnderPostmaster);
 
@@ -378,11 +379,14 @@ BootstrapModeMain(int argc, char *argv[], bool check_only)
 		Nulls[i] = false;
 	}
 
+	if (boot_yylex_init(&scanner) != 0)
+		elog(ERROR, "yylex_init() failed: %m");
+
 	/*
 	 * Process bootstrap input.
 	 */
 	StartTransactionCommand();
-	boot_yyparse();
+	boot_yyparse(scanner);
 	CommitTransactionCommand();
 
 	/*

src/include/bootstrap/bootstrap.h

@@ -55,9 +55,15 @@ extern void boot_get_type_io_data(Oid typid,
 								  Oid *typinput,
 								  Oid *typoutput);
 
-extern int	boot_yyparse(void);
+union YYSTYPE;
+#ifndef YY_TYPEDEF_YY_SCANNER_T
+#define YY_TYPEDEF_YY_SCANNER_T
+typedef void *yyscan_t;
+#endif
 
-extern int	boot_yylex(void);
-extern void boot_yyerror(const char *message) pg_attribute_noreturn();
+extern int	boot_yyparse(yyscan_t yyscanner);
+extern int	boot_yylex_init(yyscan_t *yyscannerp);
+extern int	boot_yylex(union YYSTYPE *yylval_param, yyscan_t yyscanner);
+extern void boot_yyerror(yyscan_t yyscanner, const char *message) pg_attribute_noreturn();
 
 #endif							/* BOOTSTRAP_H */