bug-868: parser confusions with braces inside ${ cmd; } form of command substitution

src/cmd/ksh93/data/options.c

@@ -39,6 +39,7 @@ const Shtable_t shtab_options[] =
 	"braceexpand",			SH_BRACEEXPAND,
 #endif
 	"noclobber",			SH_NOCLOBBER,
+	"comsub_brace_greedy",		SH_COMSUB_BRACE_GREEDY, /* bug-691: phi: */
 #if SHOPT_ESH
 	"emacs",			SH_EMACS,
 #endif

src/cmd/ksh93/include/shell.h

@@ -152,6 +152,8 @@ typedef union Shnode_u Shnode_t;
 #define SH_MULTILINE	47
 #define SH_NOBACKSLCTRL	48
 #endif
+/* bug-691: Phi: ${ list } } [;\n] } eat '}' until we get one followin [;\n] */
+#define SH_COMSUB_BRACE_GREEDY 49 /* Bug-691: */
 #define SH_LOGIN_SHELL	67
 
 #if _BLD_ksh

src/cmd/ksh93/include/shlex.h

@@ -34,6 +34,7 @@ struct _shlex_pvt_lexstate_
 {
 	char		incase;		/* 1 for case pattern, 2 after case */
 	char		intest;		/* 1 inside [[ ... ]] */
+	char		inbracecomsub;	/* 1 inside ${ ... [;\n]} bug-691: Phi:*/
 	char		testop1;	/* 1 when unary test op legal */
 	char		testop2;	/* 1 when binary test op legal */
 	char		reservok;	/* >0 for reserved word legal */

src/cmd/ksh93/sh/lex.c

@@ -312,6 +312,20 @@ int sh_lex(Lex_t* lp)
 		state = sh_lexstates[mode];
 		do {
 			n = STATE(state,c);
+                        /*
+                         * Bug-691: Phi gross hack
+                         * Greedy ${list} will treat all '}' not on a reserved
+                         * word (i.e after [;\n]) as regular char.
+                         * Non greedy close ${ on first '}' (ATT code)
+                         */
+                        if( sh_isoption(SH_COMSUB_BRACE_GREEDY) &&
+                            lp->lex.inbracecomsub && c=='}' &&
+                            ( state==sh_lexstates[ ST_BEGIN] ||
+                              state==sh_lexstates[ST_NORM]
+                            )
+                          )
+                        {  n=S_REG ; /* treat '}' as reg char */
+                        }
 			if (varnametry)
 				varnamecount += LEN;
 		} while (n == 0);
@@ -430,6 +444,10 @@ int sh_lex(Lex_t* lp)
 				lp->lex.skipword = 0;
 				/* FALLTHROUGH */
 			case S_NL:
+				/* bug-691: Phi: */
+				if( sh_isoption(SH_COMSUB_BRACE_GREEDY) )
+					lp->lex.inbracecomsub=0;
+
 				/* skip over new-lines */
 				lp->lex.last_quote = 0;
 				while(sh.inlineno++,fcget()=='\n');
@@ -453,6 +471,10 @@ int sh_lex(Lex_t* lp)
 				}
 				continue;
 			case S_OP:
+				/* bug-691: Phi: */
+				if( sh_isoption(SH_COMSUB_BRACE_GREEDY) )
+					lp->lex.inbracecomsub=0;
+
 				/* return operator token */
 				if(c=='<' || c=='>')
 				{
@@ -1001,7 +1023,7 @@ int sh_lex(Lex_t* lp)
 						continue;
 					if((n=sh_lexstates[ST_BEGIN][c])==0 || n==S_OP || n==S_NLTOK)
 					{
-						c = LBRACE;
+						c = LBRACE;      
 						goto do_comsub;
 					}
 				}
@@ -1259,6 +1281,42 @@ int sh_lex(Lex_t* lp)
 	if(!(state=lp->lexd.first))
 		state = fcfirst();
 	n = fcseek(0)-(char*)state;
+	/*
+	 * bug-691: Phi:
+	 * In the perf path (ksh interpreter) we don't enter the following if()
+	 * For shcomp we catch "set [+-]o comsub_brace_greedy" and set sh_option
+	 * accordingly, allowing compile to proceed in greedy mode if needed.
+	 */
+	if(sh.shcomp && lp->lex.reservok && n==3 &&
+	   state[0]=='s' && state[1]=='e'&& state[2]=='t' )
+	{
+		int o=0;
+		const char *p=state+3;
+		while(*p && *p!=';' && *p!='\n')
+		{
+			if(p[0]=='+' && p[1]=='o')
+			{
+				o=1;
+			}
+			if(p[0]=='-' && p[1]=='o')
+			{
+				o=2;
+			}
+			if( o && strncmp(p,"comsub_brace_greedy",19)==0)
+			{
+				if(o==1)
+				{
+					sh_offoption(SH_COMSUB_BRACE_GREEDY);
+				}
+				if(o==2)
+				{
+					sh_onoption(SH_COMSUB_BRACE_GREEDY);
+				}
+				break;
+			}
+			p++;
+		}
+	}
 	if(!lp->arg)
 		lp->arg = stkseek(sh.stk,ARGVAL);
 	if(n>0)
@@ -1281,7 +1339,12 @@ int sh_lex(Lex_t* lp)
 		if(n==LBRACT)
 			c = 0;
 		else if(n==RBRACE && lp->comsub)
-			return lp->token=n;
+		{	/* Bug-691: Phi: AT&T original non greedy path */
+			if( ! (sh_isoption(SH_COMSUB_BRACE_GREEDY) && 
+				lp->lex.inbracecomsub))
+					return lp->token=n;  
+			/* Bug-691: Phi: Greedy continue... */
+		}
 		else if(n=='~')
 			c = ARG_MAC;
 		else
@@ -1552,6 +1615,11 @@ static int comsub(Lex_t *lp, int endtok)
 	int off, messages=0, assignok=lp->assignok, csub;
 	struct _shlex_pvt_lexstate_ save = lp->lex;
 	csub = lp->comsub;
+
+	/* bug-691: Phi: */
+	if( sh_isoption(SH_COMSUB_BRACE_GREEDY) && *cp=='{' )
+		lp->lex.inbracecomsub = 1;
+
 	sh_lexopen(lp,1);
 	lp->lexd.dolparen++;
 	lp->lexd.dolparen_arithexp = endtok==LPAREN && fcpeek(1)==LPAREN;  /* $(( */
@@ -1622,13 +1690,16 @@ static int comsub(Lex_t *lp, int endtok)
 				if(endtok==LBRACE && !lp->lex.incase)
 				{
 					lp->comsub = 0;
+					/* bug-691: Phi: */
+					if(! sh_isoption(SH_COMSUB_BRACE_GREEDY))
 					count++;
 				}
 				break;
 			    case RBRACE:
 			    rbrace:
 				if(endtok==LBRACE && --count<=0)
 					goto done;
+
 				if(count==1)
 					lp->comsub = endtok==LBRACE;
 				break;

src/cmd/ksh93/sh/parse.c

@@ -492,6 +492,7 @@ Shnode_t *sh_dolparen(Lex_t* lp)
 		t = sh_cmd(lp,RPAREN,SH_NL|SH_EMPTY);
 		break;
 	    case LBRACE:
+		lp->lex.inbracecomsub=1; /* bug-691: Phi: */
 		t = sh_cmd(lp,RBRACE,SH_NL|SH_EMPTY);
 		break;
 	}

src/cmd/ksh93/tests/greedy_comsub.sh

@@ -0,0 +1,79 @@
+########################################################################
+#                                                                      #
+#              This file is part of the ksh 93u+m package              #
+#          Copyright (c) 2022-2025 Contributors to ksh 93u+m           #
+#                      and is licensed under the                       #
+#                 Eclipse Public License, Version 2.0                  #
+#                                                                      #
+#                A copy of the License is available at                 #
+#      https://www.eclipse.org/org/documents/epl-2.0/EPL-2.0.html      #
+#         (with md5 checksum 84283fa8859daf213bdda5a9f8d1be1d)         #
+#                                                                      #
+#                      Phi <[email protected]>                      #
+#                  Martijn Dekker <[email protected]>                   #
+#                                                                      #
+########################################################################
+
+
+. "${SHTESTS_COMMON:-${0%/*}/_common}"
+
+
+alias T='do_test "$LINENO"'
+
+# ======
+# Tests for greedy comsub, i.e ${ list } [;\n]}
+#                                      |      |
+#                                      |      +->   Reserved word (keyword)
+#                                      +--------> ! Reserved word (keyword)
+# https://github.com/ksh93/ksh/issues/691
+
+function do_test # $LINENO "${ list ;}" "expect"
+{ [ "$2" = "$3" ] ||
+  \err_exit "$1" "expected '$3', got '$2'"
+}
+
+
+# ======
+
+set -o | grep -q comsub_brace_greedy ||
+{ warning \
+  "This SHELL doesn't support 'set -o comsub_brace_greedy ';  skipping tests"
+  alias T=:
+}  
+set -o comsub_brace_greedy
+
+T "${ echo A ; }"                   "A"
+T "${ echo A ;}"                    "A"
+T "${ echo A;}"                     "A"
+# ======
+
+T "${ echo A ${ echo B ;} ;}"       "A B" # err_exit 
+T "${ echo A ${ echo B ;};}"        "A B"
+T "${ echo A ${ echo B;};}"         "A B"
+T "${ echo A { ${ echo B ;} C } ;}" "A { B C }"
+T "${ echo A { ${ echo B;} C } ;}"  "A { B C }"
+T "${ echo A { ${ echo B;} C};}"    "A { B C}"
+T "${ echo A ;} ${ echo B ;}"       "A B"
+T "${ echo A ;} ${ echo B;}"        "A B"
+T "${ echo A ;}${ echo B;}"         "AB"
+T "${ echo A;}${ echo B;}"          "AB"
+
+# @stephane-chazelas test cases
+# https://github.com/ksh93/ksh/issues/691
+T "${ echo {a.b}; }"                            "{a.b}"
+T "${ echo '{a,b}c' ;}"                         "{a,b}c"
+T "${ echo ${ echo {fd[0]}< /dev/null; } ;}"    ""
+
+# Variation on @stephane-chazelas
+a=1 a.c=1
+T "${ echo ${a.c}; }"                           "1"
+T "${ echo {a..c}; }"                           "a b c"
+a=x b=z
+T "${ echo {$a..$b}; }"                         "x y z"
+T "${ echo {${a}..${b}}; }"                     "x y z"
+a='x;' b='z;'
+T "${ echo {${a//;}..${b//;}}; }"               "x y z"
+
+
+# ======
+exit $((Errors<125?Errors:125))

src/cmd/ksh93/tests/shtests

@@ -376,6 +376,7 @@ do	[[ $i == *.sh ]] || i+='.sh'
 		glob.sh)	grep -c '^[[:blank:]]*test_[a-z]\{3,\}' $i ;;
 		leaks.sh)	grep -c ^TEST $i ;;
 		printf.sh)	grep -c -E '([[:blank:]]err_exit|^[[:blank:]]*T)[[:blank:]]' $i ;;
+                greedy_comsub.sh) grep -c '^T ' $i ;; # bug-691: Phi:
 		pty.sh)		grep -c 'tst ' $i ;;
 		*)		grep -c err_exit $i ;;
 		esac	)