implement ksh93 feature ${ foo;}

2012-07-30 21:37:17 +00:00
parent 8e51bb7a23
commit 9b7b7f742e
10 changed files with 146 additions and 72 deletions
--- a/check.t
+++ b/check.t
@@ -1,4 +1,4 @@
-# $MirOS: src/bin/mksh/check.t,v 1.552 2012/07/30 19:58:03 tg Exp $
+# $MirOS: src/bin/mksh/check.t,v 1.553 2012/07/30 21:37:08 tg Exp $
 # $OpenBSD: bksl-nl.t,v 1.2 2001/01/28 23:04:56 niklas Exp $
 # $OpenBSD: history.t,v 1.5 2001/01/28 23:04:56 niklas Exp $
 # $OpenBSD: read.t,v 1.3 2003/03/10 03:48:16 david Exp $
@@ -29,7 +29,7 @@
 # http://www.freebsd.org/cgi/cvsweb.cgi/src/tools/regression/bin/test/regress.sh?rev=HEAD

 expected-stdout:
-	@(#)MIRBSD KSH R40 2012/07/21
+	@(#)MIRBSD KSH R40 2012/07/30
 description:
 	Check version of shell.
 stdin:
@@ -38,7 +38,7 @@ name: KSH_VERSION
 category: shell:legacy-no
 ---
 expected-stdout:
-	@(#)LEGACY KSH R40 2012/07/21
+	@(#)LEGACY KSH R40 2012/07/30
 description:
 	Check version of legacy shell.
 stdin:
@@ -6194,21 +6194,6 @@ expected-stdout:
 expected-stderr-pattern:
 	/(Unrecognized character .... ignored at \..t4 line 1)*/
 ---
-name: utf8bom-3
-description:
-	Reading the UTF-8 BOM should enable the utf8-mode flag
-	(temporarily for COMSUBs)
-stdin:
-	"$__progname" -c ':; if [[ $- = *U* ]]; then echo 1 on; else echo 1 off; fi'
-	"$__progname" -c ':; if [[ $- = *U* ]]; then echo 2 on; else echo 2 off; fi'
-	"$__progname" -c 'if [[ $- = *U* ]]; then echo 3 on; else echo 3 off; fi; x=$(:; if [[ $- = *U* ]]; then echo 4 on; else echo 4 off; fi); echo $x; if [[ $- = *U* ]]; then echo 5 on; else echo 5 off; fi'
-expected-stdout:
-	1 off
-	2 on
-	3 off
-	4 on
-	5 off
---
 name: utf8opt-1a
 description:
 	Check that the utf8-mode flag is not set at non-interactive startup
@@ -9520,6 +9505,26 @@ expected-stdout:
 		x=$(( echo $(true >&3 ) $((1+ 2)) ) | tr u x ) 
 	} 
 ---
+name: funsub-1
+description:
+	Check that non-subenvironment command substitution works
+stdin:
+	set -e
+	foo=bar
+	echo "ob $foo ."
+	echo "${
+		echo "ib $foo :"
+		foo=baz
+		echo "ia $foo :"
+		false
+	}" .
+	echo "oa $foo ."
+expected-stdout:
+	ob bar .
+	ib bar :
+	ia baz : .
+	oa baz .
+---
 name: test-stnze-1
 description:
 	Check that the short form [ $x ] works
@@ -9613,7 +9618,6 @@ expected-stdout:
 name: event-subst-3
 description:
 	Check that '!' substitution in noninteractive mode is ignored
-category: !smksh
 file-setup: file 755 "falsetto"
 	#! /bin/sh
 	echo molto bene
--- a/dot.mkshrc
+++ b/dot.mkshrc
@@ -1,7 +1,8 @@
 # $Id$
-# $MirOS: src/bin/mksh/dot.mkshrc,v 1.69 2011/12/31 01:07:19 tg Exp $
+# $MirOS: src/bin/mksh/dot.mkshrc,v 1.70 2012/07/30 21:37:10 tg Exp $
 #-
-# Copyright (c) 2002, 2003, 2004, 2006, 2007, 2008, 2009, 2010, 2011
+# Copyright (c) 2002, 2003, 2004, 2006, 2007, 2008, 2009, 2010,
+#		2011, 2012
 #	Thorsten Glaser <tg@mirbsd.org>
 #
 # Provided that these terms and disclaimer and all copyright notices
@@ -29,7 +30,7 @@ function precmd {

 	(( e )) && print -n "$e|"
 }
-PS1=$'\001\r''$(precmd)${USER:=$(ulimit -c 0; id -un 2>/dev/null || echo \?
+PS1=$'\001\r''${ precmd;}${USER:=$(ulimit -c 0; id -un 2>/dev/null || echo \?
 	)}@${HOSTNAME%%.*}:$(local d=${PWD:-?} p=~; [[ $p = ?(*/) ]] || \
 	d=${d/#$p/~}; local m=${%d} n p=...; (( m > 0 )) || m=${#d}
 	(( m > (n = (COLUMNS/3 < 7 ? 7 : COLUMNS/3)) )) && d=${d:(-n)} || \
--- a/eval.c
+++ b/eval.c
@@ -23,7 +23,7 @@

 #include "sh.h"

-__RCSID("$MirOS: src/bin/mksh/eval.c,v 1.122 2012/07/30 17:28:21 tg Exp $");
+__RCSID("$MirOS: src/bin/mksh/eval.c,v 1.123 2012/07/30 21:37:11 tg Exp $");

 /*
 * string expansion
@@ -58,7 +58,7 @@ typedef struct Expand {
 #define IFS_NWS		2	/* have seen IFS non-white-space */

 static int varsub(Expand *, const char *, const char *, int *, int *);
-static int comsub(Expand *, const char *);
+static int comsub(Expand *, const char *, int);
 static char *trimsub(char *, char *, int);
 static void glob(char *, XPtrV *, int);
 static void globit(XString *, char **, char *, XPtrV *, int);
@@ -278,17 +278,27 @@ expand(const char *cp,	/* input word */
 				quote = st->quotew;
 				continue;
 			case COMSUB:
+			case FUNSUB:
 				tilde_ok = 0;
 				if (f & DONTRUNCOMMAND) {
 					word = IFS_WORD;
-					*dp++ = '$'; *dp++ = '(';
+					*dp++ = '$';
+					if (c == FUNSUB) {
+						*dp++ = '{';
+						*dp++ = ' ';
+					} else
+						*dp++ = '(';
 					while (*sp != '\0') {
 						Xcheck(ds, dp);
 						*dp++ = *sp++;
 					}
+					if (c == FUNSUB) {
+						*dp++ = ';';
+						*dp++ = '}';
+					} else
 						*dp++ = ')';
 				} else {
-					type = comsub(&x, sp);
+					type = comsub(&x, sp, c);
 					if (type == XCOM && (f&DOBLANK))
 						doblank++;
 					sp = strnul(sp) + 1;
@@ -611,8 +621,8 @@ expand(const char *cp,	/* input word */
 					case '#':
 					case '%':
 						/* ! DOBLANK,DOBRACE,DOTILDE */
-						f = DOPAT | (f&DONTRUNCOMMAND) |
-						    DOTEMP;
+						f = (f & DONTRUNCOMMAND) |
+						    DOPAT | DOTEMP;
 						st->quotew = quote = 0;
 						/*
 						 * Prepend open pattern (so |
@@ -1272,7 +1282,7 @@ varsub(Expand *xp, const char *sp, const char *word,
 * Run the command in $(...) and read its output.
 */
 static int
-comsub(Expand *xp, const char *cp)
+comsub(Expand *xp, const char *cp, int fn)
 {
 	Source *s, *sold;
 	struct op *t;
@@ -1286,10 +1296,15 @@ comsub(Expand *xp, const char *cp)
 	afree(s, ATEMP);
 	source = sold;

+	UTFMODE = old_utfmode;
+
 	if (t == NULL)
 		return (XBASE);

-	if (t != NULL && t->type == TCOM &&
+	/* no waitlast() unless specifically enabled later */
+	xp->split = false;
+
+	if (t->type == TCOM &&
 	    *t->args == NULL && *t->vars == NULL && t->ioact != NULL) {
 		/* $(<file) */
 		struct ioword *io = *t->ioact;
@@ -1302,10 +1317,9 @@ comsub(Expand *xp, const char *cp)
 			SHF_MAPHI|SHF_CLEXEC);
 		if (shf == NULL)
 			errorf("%s: %s %s", name, "can't open", "$() input");
-		/* no waitlast() */
-		xp->split = false;
 	} else {
 		int ofd1, pv[2];
+
 		openpipe(pv);
 		shf = shf_fdopen(pv[0], SHF_RD, NULL);
 		ofd1 = savefd(1);
@@ -1313,14 +1327,16 @@ comsub(Expand *xp, const char *cp)
 			ksh_dup2(pv[1], 1, false);
 			close(pv[1]);
 		}
-		execute(t, XFORK|XXCOM|XPIPEO, NULL);
+		execute(t, XXCOM | XPIPEO |
+		    (fn == FUNSUB ? XERROK : XFORK), NULL);
 		restfd(1, ofd1);
+		if (fn != FUNSUB) {
 			startlast();
 			/* waitlast() */
 			xp->split = true;
 		}
+	}

-	UTFMODE = old_utfmode;
 	xp->u.shf = shf;
 	return (XCOM);
 }
@@ -1328,7 +1344,6 @@ comsub(Expand *xp, const char *cp)
 /*
 * perform #pattern and %pattern substitution in ${}
 */
-
 static char *
 trimsub(char *str, char *pat, int how)
 {
--- a/lex.c
+++ b/lex.c
@@ -23,7 +23,7 @@

 #include "sh.h"

-__RCSID("$MirOS: src/bin/mksh/lex.c,v 1.165 2012/07/01 15:54:55 tg Exp $");
+__RCSID("$MirOS: src/bin/mksh/lex.c,v 1.166 2012/07/30 21:37:12 tg Exp $");

 /*
 * states while lexing word
@@ -109,7 +109,7 @@ void yyskiputf8bom(void);
 static int backslash_skip;
 static int ignore_backslash_newline;
 static struct sretrace_info *retrace_info;
-short subshell_nesting_level = 0;
+int subshell_nesting_type = 0;

 /* optimised getsc_bn() */
 #define o_getsc()	(*source->str != '\0' && *source->str != '\\' && \
@@ -262,6 +262,11 @@ yylex(int cf)
 	while (!((c = getsc()) == 0 ||
 	    ((state == SBASE || state == SHEREDELIM || state == SHERESTRING) &&
 	    ctype(c, C_LEX1)))) {
+		if (state == SBASE &&
+		    subshell_nesting_type == /*{*/ '}' &&
+		    c == /*{*/ '}')
+			/* possibly end ${ :;} */
+			break;
 accept_nonword:
 		Xcheck(ws, wp);
 		switch (state) {
@@ -395,14 +400,27 @@ yylex(int cf)
 					} else {
 						ungetsc(c);
 subst_command:
-						sp = yyrecursive();
+						c = COMSUB;
+ subst_command2:
+						sp = yyrecursive(c);
 						cz = strlen(sp) + 1;
 						XcheckN(ws, wp, cz);
-						*wp++ = COMSUB;
+						*wp++ = c;
 						memcpy(wp, sp, cz);
 						wp += cz;
 					}
 				} else if (c == '{') /*}*/ {
+					c = getsc();
+					if (ctype(c, C_IFSWS)) {
+						/*
+						 * non-subenvironment
+						 * "command" substitution
+						 */
+						c = FUNSUB;
+						goto subst_command2;
+					}
+					ungetsc(c);
+
 					*wp++ = OSUBST;
 					*wp++ = '{'; /*}*/
 					wp = get_brace_var(&ws, wp);
@@ -1171,7 +1189,7 @@ readhere(struct ioword *iop)
 		/* end of here document marker, what to do? */
 		switch (c) {
 		case /*(*/ ')':
-			if (!subshell_nesting_level)
+			if (!subshell_nesting_type)
 				/*-
 				 * not allowed outside $(...) or (...)
 				 * => mismatch
--- a/misc.c
+++ b/misc.c
@@ -30,7 +30,7 @@
 #include <grp.h>
 #endif

-__RCSID("$MirOS: src/bin/mksh/misc.c,v 1.196 2012/07/20 21:18:45 tg Exp $");
+__RCSID("$MirOS: src/bin/mksh/misc.c,v 1.197 2012/07/30 21:37:13 tg Exp $");

 #define KSH_CHVT_FLAG
 #ifdef MKSH_SMALL
@@ -101,9 +101,9 @@ initctypes(void)
 	chtypes['_'] |= C_ALPHA;
 	setctypes("0123456789", C_DIGIT);
 	/* \0 added automatically */
-	setctypes(" \t\n|&;<>()", C_LEX1);
+	setctypes(TC_LEX1, C_LEX1);
 	setctypes("*@#!$-?", C_VAR1);
-	setctypes(" \t\n", C_IFSWS);
+	setctypes(TC_IFSWS, C_IFSWS);
 	setctypes("=-+?", C_SUBOP1);
 	setctypes("\t\n \"#$&'()*;<=>?[\\]`|", C_QUOTE);
 }
--- a/mksh.1
+++ b/mksh.1
@@ -1,4 +1,4 @@
-.\" $MirOS: src/bin/mksh/mksh.1,v 1.290 2012/07/25 19:40:40 tg Exp $
+.\" $MirOS: src/bin/mksh/mksh.1,v 1.291 2012/07/30 21:37:13 tg Exp $
 .\" $OpenBSD: ksh.1,v 1.144 2012/07/08 08:13:20 guenther Exp $
 .\"-
 .\" Copyright © 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009,
@@ -74,7 +74,7 @@
 .\" with -mandoc, it might implement .Mx itself, but we want to
 .\" use our own definition. And .Dd must come *first*, always.
 .\"
-.Dd $Mdocdate: July 25 2012 $
+.Dd $Mdocdate: July 30 2012 $
 .\"
 .\" Check which macro package we use, and do other -mdoc setup.
 .\"
@@ -1145,9 +1145,15 @@ or
 command substitutions take the form
 .Pf $( Ns Ar command Ns \&)
 or (deprecated)
-.Pf \` Ns Ar command Ns \` ;
+.Pf \` Ns Ar command Ns \`
+or (executed in the current environment)
+.Pf ${\ \& Ar command Ns \&;}
+and strip trailing newlines;
 and arithmetic substitutions take the form
 .Pf $(( Ns Ar expression Ns )) .
+Parsing the current-environment command substitution requires a space,
+tab or newline after the opening brace and that the closing brace be
+recognised as a keyword (i.e. is preceded by a newline or semicolon).
 .Pp
 If a substitution appears outside of double quotes, the results of the
 substitution are generally subject to word or field splitting according to
@@ -1225,6 +1231,8 @@ A command substitution is replaced by the output generated by the specified
 command which is run in a subshell.
 For
 .Pf $( Ns Ar command Ns \&)
+and
+.Pf ${\ \& Ar command Ns \&;}
 substitutions, normal quoting rules are used when
 .Ar command
 is parsed; however, for the deprecated
@@ -1963,7 +1971,7 @@ x=$(print \e\e001)
 PS1="$x$(print \e\er)$x$(tput smso)$x\e$PWD$x$(tput rmso)$x\*(Gt "
 .Ed
 .Pp
-Due to pressure from David G. Korn,
+Due to a strong suggestion from David G. Korn,
 .Nm
 now also supports the following form:
 .Bd -literal -offset indent
--- a/sh.h
+++ b/sh.h
@@ -157,9 +157,9 @@
 #endif

 #ifdef EXTERN
-__RCSID("$MirOS: src/bin/mksh/sh.h,v 1.575 2012/07/30 17:04:30 tg Exp $");
+__RCSID("$MirOS: src/bin/mksh/sh.h,v 1.576 2012/07/30 21:37:15 tg Exp $");
 #endif
-#define MKSH_VERSION "R40 2012/07/21"
+#define MKSH_VERSION "R40 2012/07/30"

 /* arithmetic types: C implementation */
 #if !HAVE_CAN_INTTYPES
@@ -765,6 +765,8 @@ EXTERN const char Tgbuiltin[] E_INIT("=builtin");
 #define Tbuiltin	(Tgbuiltin + 1)		/* "builtin" */
 EXTERN const char T_function[] E_INIT(" function");
 #define Tfunction	(T_function + 1)	/* "function" */
+EXTERN const char TC_LEX1[] E_INIT("|&;<>() \t\n");
+#define TC_IFSWS	(TC_LEX1 + 7)		/* space tab newline */

 typedef uint8_t Temp_type;
 /* expanded heredoc */
@@ -1281,6 +1283,7 @@ struct op {
 #define SPAT	10	/* separate pattern: | */
 #define CPAT	11	/* close pattern: ) */
 #define ADELIM	12	/* arbitrary delimiter: ${foo:2:3} ${foo/bar/baz} */
+#define FUNSUB	14	/* ${ foo;} substitution (NUL terminated) */

 /*
 * IO redirection
@@ -1903,7 +1906,7 @@ ssize_t shf_vfprintf(struct shf *, const char *, va_list)
 void initkeywords(void);
 struct op *compile(Source *, bool);
 bool parse_usec(const char *, struct timeval *);
-char *yyrecursive(void);
+char *yyrecursive(int);
 /* tree.c */
 void fptreef(struct shf *, int, const char *, ...);
 char *snptreef(char *, ssize_t, const char *, ...);
--- a/syn.c
+++ b/syn.c
@@ -23,9 +23,9 @@

 #include "sh.h"

-__RCSID("$MirOS: src/bin/mksh/syn.c,v 1.78 2012/07/30 17:04:31 tg Exp $");
+__RCSID("$MirOS: src/bin/mksh/syn.c,v 1.79 2012/07/30 21:37:16 tg Exp $");

-extern short subshell_nesting_level;
+extern int subshell_nesting_type;
 extern void yyskiputf8bom(void);

 struct nesting_state {
@@ -363,12 +363,15 @@ get_command(int cf)
 Leave:
 		break;

-	case '(':
+	case '(': /*)*/ {
+		int subshell_nesting_type_saved;
 Subshell:
-		++subshell_nesting_level;
+		subshell_nesting_type_saved = subshell_nesting_type;
+		subshell_nesting_type = ')';
 		t = nested(TPAREN, '(', ')');
-		--subshell_nesting_level;
+		subshell_nesting_type = subshell_nesting_type_saved;
 		break;
+	    }

 	case '{': /*}*/
 		t = nested(TBRACE, '{', '}');
@@ -1115,16 +1118,29 @@ parse_usec(const char *s, struct timeval *tv)
 * a COMSUB recursively using the main shell parser and lexer
 */
 char *
-yyrecursive(void)
+yyrecursive(int subtype)
 {
 	struct op *t;
 	char *cp;
 	bool old_reject;
-	int old_symbol, old_salias;
+	int old_symbol, old_salias, old_nesting_type;
 	struct ioword **old_herep;
+	int stok, etok;
+
+	switch (subtype) {
+	case FUNSUB:
+		stok = '{';
+		etok = '}';
+		break;
+	case COMSUB:
+	default:
+		stok = '(';
+		etok = ')';
+	}

 	/* tell the lexer to accept a closing parenthesis as EOD */
-	++subshell_nesting_level;
+	old_nesting_type = subshell_nesting_type;
+	subshell_nesting_type = etok;

 	/* push reject state, parse recursively, pop reject state */
 	old_reject = reject;
@@ -1134,7 +1150,7 @@ yyrecursive(void)
 	old_salias = sALIAS;
 	sALIAS = 0;
 	/* we use TPAREN as a helper container here */
-	t = nested(TPAREN, '(', ')');
+	t = nested(TPAREN, stok, etok);
 	sALIAS = old_salias;
 	herep = old_herep;
 	reject = old_reject;
@@ -1144,6 +1160,6 @@ yyrecursive(void)
 	cp = snptreef(NULL, 0, "%T", t->left);
 	tfree(t, ATEMP);

-	--subshell_nesting_level;
+	subshell_nesting_type = old_nesting_type;
 	return (cp);
 }
--- a/tree.c
+++ b/tree.c
@@ -23,7 +23,7 @@

 #include "sh.h"

-__RCSID("$MirOS: src/bin/mksh/tree.c,v 1.62 2012/07/30 19:58:05 tg Exp $");
+__RCSID("$MirOS: src/bin/mksh/tree.c,v 1.63 2012/07/30 21:37:17 tg Exp $");

 #define INDENT	8

@@ -292,6 +292,7 @@ static const char *
 wdvarput(struct shf *shf, const char *wp, int quotelevel, int opmode)
 {
 	int c;
+	const char *cs;

 	/*-
 	 * problems:
@@ -335,16 +336,20 @@ wdvarput(struct shf *shf, const char *wp, int quotelevel, int opmode)
 		}
 		case COMSUB:
 			shf_puts("$(", shf);
+			cs = ")";
+ pSUB:
 			while ((c = *wp++) != 0)
 				shf_putc(c, shf);
-			shf_putc(')', shf);
+			shf_puts(cs, shf);
 			break;
+		case FUNSUB:
+			shf_puts("${ ", shf);
+			cs = ";}";
+			goto pSUB;
 		case EXPRSUB:
 			shf_puts("$((", shf);
-			while ((c = *wp++) != 0)
-				shf_putc(c, shf);
-			shf_puts("))", shf);
-			break;
+			cs = "))";
+			goto pSUB;
 		case OQUOTE:
 			if (opmode & WDS_TPUTS) {
 				quotelevel++;
@@ -580,6 +585,7 @@ wdscan(const char *wp, int c)
 			wp++;
 			break;
 		case COMSUB:
+		case FUNSUB:
 		case EXPRSUB:
 			while (*wp++ != 0)
 				;
@@ -819,6 +825,9 @@ dumpwdvar_i(struct shf *shf, const char *wp, int quotelevel)
 closeandout:
 			shf_putc('>', shf);
 			break;
+		case FUNSUB:
+			shf_puts("FUNSUB<", shf);
+			goto dumpsub;
 		case EXPRSUB:
 			shf_puts("EXPRSUB<", shf);
 			goto dumpsub;
--- a/var.c
+++ b/var.c
@@ -27,7 +27,7 @@
 #include <sys/sysctl.h>
 #endif

-__RCSID("$MirOS: src/bin/mksh/var.c,v 1.152 2012/07/01 15:38:09 tg Exp $");
+__RCSID("$MirOS: src/bin/mksh/var.c,v 1.153 2012/07/30 21:37:17 tg Exp $");

 /*-
 * Variables
@@ -1254,7 +1254,7 @@ unsetspec(struct tbl *vp)
 		flushcom(true);
 		break;
 	case V_IFS:
-		setctypes(" \t\n", C_IFS);
+		setctypes(TC_IFSWS, C_IFS);
 		ifs0 = ' ';
 		break;
 	case V_TMPDIR: