New feature: 'typeset -g' as in bash 4.2+

typeset -g allows directly manipulating the attributes of variables at the global level from any context. This feature already exists on bash 4.2 and later. mksh (R50+), yash and zsh have this flag as well, but it's slightly different: it ignores the current local scope, but a parent local scope from a calling function may still be used -- whereas on bash, '-g' always refers to the global scope. Since ksh93 uses static scoping (see III.Q28 at <http://kornshell.com/doc/faq.html>), only the bash behaviour makes sense here. Note that the implementation needs to be done both in nv_setlist() (name.c) and in b_typeset() (typeset.c) because assignments are executed before the typeset built-in itself. Hence also the pre-parsing of typeset options in sh_exec(). src/cmd/ksh93/include/nval.h: - Add new NV_GLOBAL bit flag, using a previously unused bit that still falls within the 32-bit integer range. src/cmd/ksh93/sh/xec.c: sh_exec(): - When pre-parsing typeset flags, make -g pass the NV_GLOBAL flag to the nv_setlist() call that processes shell assignments prior to running the command. src/cmd/ksh93/sh/name.c: nv_setlist(): - When the NV_GLOBAL bit flag is passed, save the current variable tree pointer (sh.var_tree) as well as the current namespace (sh.namespace) and temporarily set the former to the global variable tree (sh.var_base) and the latter to NULL. This makes assignments global and ignores namesapces. src/cmd/ksh93/bltins/typeset.c: - b_typeset(): - Use NV_GLOBAL bit flag for -g. - Allow combining -n with -g, permitting 'typeset -gn var' or 'nameref -g var' to create a global nameref from a function. - Do not allow a nonsensical use of -g when using nested typeset to define member variables of 'typeset -T' types. (Type method functions can still use it as normal.) - setall(): - If NV_GLOBAL is passed, use sh.var_base and deactivate sh.namespace as in nv_setlist(). This allows attributes to be set correctly for global variables. src/cmd/ksh93/tests/{functions,namespace}.sh: - Add regression tests based on reproducers for problems found by @hyenias in preliminary versions of this feature. Resolves: https://github.com/ksh93/ksh/issues/479
2025-03-09 15:50:02 +00:00 · 2022-06-01 17:14:51 +01:00 · 2022-06-01 17:14:51 +01:00 · ea300089a1
commit ea300089a1
parent f73b8617dd
11 changed files with 167 additions and 8 deletions
--- a/3
+++ b/3
@ -56,6 +56,9 @@ New features in built-in commands:
  Note that to use these options the operating system must support the
  corresponding resource limit.

+- 'typeset' has a new '-g' flag that forces variables to be created or
+  modified at the global scope regardless of context, as on bash 4.2+.
+
 New command line editor features:

 - Various keys on extended PC keyboards are now handled as expected in the
--- a/8
+++ b/8
@ -3,6 +3,14 @@ For full details, see the git log at: https://github.com/ksh93/ksh/tree/1.0

 Any uppercase BUG_* names are modernish shell bug IDs.

+2022-06-01:
+
+- New 'typeset' feature inspired by bash (and zsh and mksh): the -g flag
+  causes the specified variable(s) to be created globally, ignoring the
+  local scope of ksh functions or name spaces. For example, this allows
+  creating a global array or numeric variable from a ksh function.
+  Thanks to @ormaaj for the feature suggestion and to @hyenias for testing.
+
 2022-05-29:

 - Fixed a bug causing an inconsistent state after a special built-in command
--- a/src/cmd/ksh93/bltins/typeset.c
+++ b/src/cmd/ksh93/bltins/typeset.c
@ -424,6 +424,9 @@ int    b_typeset(int argc,register char *argv[],Shbltin_t *context)
 				flag &= ~NV_VARNAME;
 				flag |= (NV_EXPORT|NV_IDENT);
 				break;
+			case 'g':
+				flag |= NV_GLOBAL;
+				break;
 			case ':':
 				errormsg(SH_DICT,2, "%s", opt_info.arg);
 				break;
@ -458,9 +461,9 @@ endargs:
 		errormsg(SH_DICT,2,e_optincompat1,"-m");
 		error_info.errors++;
 	}
-	if((flag&NV_REF) && (flag&~(NV_REF|NV_IDENT|NV_ASSIGN)))
+	if((flag&NV_REF) && (flag&~(NV_REF|NV_IDENT|NV_ASSIGN|NV_GLOBAL)))
 	{
-		errormsg(SH_DICT,2,e_optincompat1,"-n");
+		errormsg(SH_DICT,2,e_optincompat2,"-n","other options except -g");
 		error_info.errors++;
 	}
 	if((flag&NV_TYPE) && (flag&~(NV_TYPE|NV_VARNAME|NV_ASSIGN)))
@ -489,6 +492,11 @@ endargs:
 		errormsg(SH_DICT,ERROR_exit(2),"option argument cannot be greater than %d",SHRT_MAX);
 		UNREACHABLE();
 	}
+	if((flag&NV_GLOBAL) && sh.mktype)
+	{
+		errormsg(SH_DICT,ERROR_exit(2),"-g: type members cannot be global");
+		UNREACHABLE();
+	}
 	if(isfloat)
 		flag |= NV_DOUBLE;
 	if(sflag)
@ -634,6 +642,17 @@ static int     setall(char **argv,register int flag,Dt_t *troot,struct tdata *tp
 	char *last = 0;
 	int nvflags=(flag&(NV_ARRAY|NV_NOARRAY|NV_VARNAME|NV_IDENT|NV_ASSIGN|NV_STATIC|NV_MOVE));
 	int r=0, ref=0, comvar=(flag&NV_COMVAR),iarray=(flag&NV_IARRAY);
+	Dt_t *save_vartree;
+	Namval_t *save_namespace;
+	if(flag&NV_GLOBAL)
+	{
+		save_vartree = sh.var_tree;
+		troot = sh.var_tree = sh.var_base;
+#if SHOPT_NAMESPACE
+		save_namespace = sh.namespace;
+		sh.namespace = NIL(Namval_t*);
+#endif
+	}
 	if(!sh.prefix)
 	{
 		if(!tp->pflag)
@ -1005,6 +1024,13 @@ static int     setall(char **argv,register int flag,Dt_t *troot,struct tdata *tp
 		if(r==0)
 			r = 1;  /* ensure the exit status is at least 1 */
 	}
+	if(flag&NV_GLOBAL)
+	{
+		sh.var_tree = save_vartree;
+#if SHOPT_NAMESPACE
+		sh.namespace = save_namespace;
+#endif
+	}
 	return(r);
 }

--- a/src/cmd/ksh93/data/builtins.c
+++ b/src/cmd/ksh93/data/builtins.c
@ -1776,7 +1776,7 @@ const char sh_opttrap[] =
 ;

 const char sh_opttypeset[] =
-"+[-1c?\n@(#)$Id: typeset (ksh 93u+m) 2021-12-17 $\n]"
+"+[-1c?\n@(#)$Id: typeset (ksh 93u+m) 2022-06-01 $\n]"
 "[--catalog?" SH_DICT "]"
 "[+NAME?typeset - declare or display variables with attributes]"
 "[+DESCRIPTION?Without the \b-f\b option, \btypeset\b sets, unsets, "
@ -1833,6 +1833,7 @@ const char sh_opttypeset[] =
 	"encoding of the data. This option can be used with \b-Z\b to "
 	"specify fixed-size fields.]"
 "[f?Each of the options and \aname\as refers to a function.]"
+"[g?Forces variables to be created or modified at the global scope.]"
 "[i]#?[base:=10?An integer. \abase\a represents the arithmetic base "
 	"from 2 to 64.]"
 "[l?Without \b-i\b, sets character mapping to \btolower\b. When used "
@ -1896,7 +1897,7 @@ const char sh_opttypeset[] =
 "\n[name[=value]...]\n"
 " -f [name...]\n"
 " -m [name=name...]\n"
-" -n [name=name...]\n"
+" -n [-g] [name=name...]\n"
 " -T [tname[=(type definition)]...]\n"
 "\n"
 "[+EXIT STATUS?]{"
--- a/src/cmd/ksh93/include/nval.h
+++ b/src/cmd/ksh93/include/nval.h
@ -184,6 +184,7 @@ struct Namval
 #define NV_NOSCOPE	0x80000		/* look only in current scope */
 #define NV_NOFAIL	0x100000	/* return 0 on failure, no msg */
 #define NV_NODISC	NV_IDENT	/* ignore disciplines */
+#define NV_GLOBAL	0x20000000	/* create global variable, ignoring local scope */

 #define NV_FUNCT	NV_IDENT	/* option for nv_create */
 #define NV_BLTINOPT	NV_ZFILL	/* mark builtins in libcmd */
--- a/src/cmd/ksh93/include/version.h
+++ b/src/cmd/ksh93/include/version.h
@ -21,7 +21,7 @@

 #define SH_RELEASE_FORK	"93u+m"		/* only change if you develop a new ksh93 fork */
 #define SH_RELEASE_SVER	"1.0.0-beta.2"	/* semantic version number: https://semver.org */
-#define SH_RELEASE_DATE	"2022-05-29"	/* must be in this format for $((.sh.version)) */
+#define SH_RELEASE_DATE	"2022-06-01"	/* must be in this format for $((.sh.version)) */
 #define SH_RELEASE_CPYR	"(c) 2020-2022 Contributors to ksh " SH_RELEASE_FORK

 /* Scripts sometimes field-split ${.sh.version}, so don't change amount of whitespace. */
--- a/src/cmd/ksh93/sh.1
+++ b/src/cmd/ksh93/sh.1
@ -8323,6 +8323,19 @@ format
 can be used to output the actual data in this buffer instead
 of the base64 encoding of the data.
 .TP
+.B \-g
+Forces variables to be created or modified at the global scope,
+even when
+.B typeset
+is executed in a function defined by the
+.B function
+.I name
+syntax (see
+.I Functions\^
+above) or in a name space (see
+.I Name Spaces\^
+above).
+.TP
 .B \-h
 Used within type definitions to add information when generating
 information about the subvariable on the man page.
@ -8359,7 +8372,7 @@ Equivalent to
 .B "\-M tolower" .
 .TP
 .B \-m
-moves or renames the variable.
+Moves or renames the variable.
 The value is the name of a variable whose value will be moved to
 .IR vname\^ .
 The original variable will be unset.
@ -8373,7 +8386,8 @@ defined by the value of variable
 .IR vname .
 This is usually used to reference a variable inside
 a function whose name has been passed as an argument.
-Cannot be used with any other options.
+Cannot be used with other options except
+.BR \-g .
 .TP
 .B \-p
 The name, attributes and values for the given
--- a/src/cmd/ksh93/sh/name.c
+++ b/src/cmd/ksh93/sh/name.c
@ -273,6 +273,19 @@ void nv_setlist(register struct argnod *arg,register int flags, Namval_t *typ)
 	struct Namref	nr;
 	int		maketype = flags&NV_TYPE;  /* make a 'typeset -T' type definition command */
 	struct sh_type	shtp;
+	Dt_t		*save_vartree;
+#if SHOPT_NAMESPACE
+	Namval_t	*save_namespace;
+#endif
+	if(flags&NV_GLOBAL)
+	{
+		save_vartree = sh.var_tree;
+		sh.var_tree = sh.var_base;
+#if SHOPT_NAMESPACE
+		save_namespace = sh.namespace;
+		sh.namespace = NIL(Namval_t*);
+#endif
+	}
 	if(maketype)
 	{
 		shtp.previous = sh.mktype;
@ -646,6 +659,13 @@ void nv_setlist(register struct argnod *arg,register int flags, Namval_t *typ)
 		}
 		/* continue loop */
 	}
+	if(flags&NV_GLOBAL)
+	{
+		sh.var_tree = save_vartree;
+#if SHOPT_NAMESPACE
+		sh.namespace = save_namespace;
+#endif
+	}
 }

 /*
--- a/src/cmd/ksh93/sh/xec.c
+++ b/src/cmd/ksh93/sh/xec.c
@ -1068,6 +1068,8 @@ int sh_exec(register const Shnode_t *t, int flags)
 							flgs |= NV_STATIC;
 						if(checkopt(com,'m'))
 							flgs |= NV_MOVE;
+						if(checkopt(com,'g'))
+							flgs |= NV_GLOBAL;
 						if(np==SYSNAMEREF || checkopt(com,'n'))
 							flgs |= NV_NOREF;
 						else if(argn>=3 && checkopt(com,'T'))
--- a/src/cmd/ksh93/tests/functions.sh
+++ b/src/cmd/ksh93/tests/functions.sh
@ -1341,5 +1341,44 @@ got=$( { "$SHELL" "$tmp/crash_rhbz1117404.ksh"; } 2>&1)
 ((!(e = $?))) || err_exit 'crash while handling function-local trap' \
 	"(got status $e$( ((e>128)) && print -n /SIG && kill -l "$e"), $(printf %q "$got"))"

+# ======
+# 'typeset -g' should cause the active function scope to be ignored.
+# https://github.com/ksh93/ksh/issues/479#issuecomment-1140523291
+if	! command typeset -g x 2>/dev/null
+then
+	warning "shell does not have 'typeset -g'; skipping those tests"
+else
+	unset x arr a b c
+	# must specify the length for -X as the default is -X 24 on 32-bit systems and -X 32 on 64-bit systems
+	typeset -A exp=(
+		['x=123; function f { typeset -g x=456; }; function g { typeset x=789; f; }; g; typeset -p x']='x=456'
+		['x=123; function f { typeset -g -i x; }; f; typeset -p x']='typeset -i x=123'
+		['x=123; function f { typeset -g -F x; }; f; typeset -p x']='typeset -F x=123.0000000000'
+		['x=123; function f { typeset -g -E x; }; f; typeset -p x']='typeset -E x=123'
+		['x=123; function f { typeset -g -X24 x; }; f; typeset -p x']='typeset -X 24 x=0x1.ec0000000000000000000000p+6'
+		['x=aBc; function f { typeset -g -u x; }; f; typeset -p x']='typeset -u x=ABC'
+		['x=aBc; function f { typeset -g -l x; }; f; typeset -p x']='typeset -l x=abc'
+		['x=aBc; function f { typeset -g -L x; }; f; typeset -p x']='typeset -L 3 x=aBc'
+		['x=aBc; function f { typeset -g -R x; }; f; typeset -p x']='typeset -R 3 x=aBc'
+		['x=aBc; function f { typeset -g -Z x; }; f; typeset -p x']='typeset -Z 3 -R 3 x=aBc'
+		['x=aBc; function f { typeset -g -L2 x; }; f; typeset -p x']='typeset -L 2 x=aB'
+		['x=aBc; function f { typeset -g -R2 x; }; f; typeset -p x']='typeset -R 2 x=Bc'
+		['x=aBc; function f { typeset -g -Z2 x; }; f; typeset -p x']='typeset -Z 2 -R 2 x=Bc'
+		['x=8; function f { typeset -g -i2 x; }; f; typeset -p x']='typeset -i 2 x=2#1000'
+		['x=8; function f { typeset -g -i8 x; }; f; typeset -p x']='typeset -i 8 x=8#10'
+		['arr=(a b c); function f { typeset -g -i arr[1]=(1 2 3); }; f; typeset -p arr']='typeset -a -i arr=(0 (1 2 3) 0)'
+		['arr=(a b c); function f { typeset -g -F arr[1]=(1 2 3); }; f; typeset -p arr']='typeset -a -F arr=(0.0000000000 (1.0000000000 2.0000000000 3.0000000000) 0.0000000000)'
+		['arr=(a b c); function f { typeset -g -E arr[1]=(1 2 3); }; f; typeset -p arr']='typeset -a -E arr=(0 (1 2 3) 0)'
+		['arr=(a b c); function f { typeset -g -X24 arr[1]=(1 2 3); }; f; typeset -p arr']='typeset -a -X 24 arr=(0x0.000000000000000000000000p+0 (0x1.000000000000000000000000p+0 0x1.000000000000000000000000p+1 0x1.800000000000000000000000p+1) 0x0.000000000000000000000000p+0)'
+	)
+	for cmd in "${!exp[@]}"
+	do
+		got=$(set +x; eval "$cmd" 2>&1)
+		[[ $got == "${exp[$cmd]}" ]] || err_exit "typeset -g in $(printf %q "$cmd") failed to activate global" \
+			"scope from ksh function (expected $(printf %q "${exp[$cmd]}"), got $(printf %q "$got"))"
+	done
+	unset exp
+fi
+
 # ======
 exit $((Errors<125?Errors:125))
--- a/src/cmd/ksh93/tests/namespace.sh
+++ b/src/cmd/ksh93/tests/namespace.sh
@ -138,9 +138,54 @@ set +o allexport
 # is not executed -- the zillionth bug with virtual subshells >:-/
 exp=$'123 456\n789 456'
 got=$(ulimit -t unlimited 2>/dev/null  # workaround: force fork
-	trap 'echo $x ${.test.x}; x=789; echo $x ${.test.x}' EXIT; x=123; namespace test { x=456; set -Q 2>/dev/null; })
+	trap 'echo $x ${.test.x}; x=789; echo $x ${.test.x}' EXIT
+	x=123
+	namespace test
+	{
+		x=456
+		set --bad_option 2>/dev/null
+	}
+)
 [[ $got == "$exp" ]] || err_exit 'Parent scope not restored after special builtin throws error in namespace' \
 	"(expected $(printf %q "$exp"), got $(printf %q "$got"))"

+# ======
+# 'typeset -g' should cause the active name space to be ignored.
+# https://github.com/ksh93/ksh/issues/479#issuecomment-1140523291
+if	! command typeset -g x 2>/dev/null
+then
+	warning "shell does not have 'typeset -g'; skipping those tests"
+else
+	unset x arr a b c
+	# must specify the length for -X as the default is -X 24 on 32-bit systems and -X 32 on 64-bit systems
+	typeset -A exp=(
+		['x=123; namespace test { typeset -g -i x; }; typeset -p x']='typeset -i x=123'
+		['x=123; namespace test { typeset -g -F x; }; typeset -p x']='typeset -F x=123.0000000000'
+		['x=123; namespace test { typeset -g -E x; }; typeset -p x']='typeset -E x=123'
+		['x=123; namespace test { typeset -g -X24 x; }; typeset -p x']='typeset -X 24 x=0x1.ec0000000000000000000000p+6'
+		['x=aBc; namespace test { typeset -g -u x; }; typeset -p x']='typeset -u x=ABC'
+		['x=aBc; namespace test { typeset -g -l x; }; typeset -p x']='typeset -l x=abc'
+		['x=aBc; namespace test { typeset -g -L x; }; typeset -p x']='typeset -L 3 x=aBc'
+		['x=aBc; namespace test { typeset -g -R x; }; typeset -p x']='typeset -R 3 x=aBc'
+		['x=aBc; namespace test { typeset -g -Z x; }; typeset -p x']='typeset -Z 3 -R 3 x=aBc'
+		['x=aBc; namespace test { typeset -g -L2 x; }; typeset -p x']='typeset -L 2 x=aB'
+		['x=aBc; namespace test { typeset -g -R2 x; }; typeset -p x']='typeset -R 2 x=Bc'
+		['x=aBc; namespace test { typeset -g -Z2 x; }; typeset -p x']='typeset -Z 2 -R 2 x=Bc'
+		['x=8; namespace test { typeset -g -i2 x; }; typeset -p x']='typeset -i 2 x=2#1000'
+		['x=8; namespace test { typeset -g -i8 x; }; typeset -p x']='typeset -i 8 x=8#10'
+		['arr=(a b c); namespace test { typeset -g -i arr[1]=(1 2 3); }; typeset -p arr']='typeset -a -i arr=(0 (1 2 3) 0)'
+		['arr=(a b c); namespace test { typeset -g -F arr[1]=(1 2 3); }; typeset -p arr']='typeset -a -F arr=(0.0000000000 (1.0000000000 2.0000000000 3.0000000000) 0.0000000000)'
+		['arr=(a b c); namespace test { typeset -g -E arr[1]=(1 2 3); }; typeset -p arr']='typeset -a -E arr=(0 (1 2 3) 0)'
+		['arr=(a b c); namespace test { typeset -g -X24 arr[1]=(1 2 3); }; typeset -p arr']='typeset -a -X 24 arr=(0x0.000000000000000000000000p+0 (0x1.000000000000000000000000p+0 0x1.000000000000000000000000p+1 0x1.800000000000000000000000p+1) 0x0.000000000000000000000000p+0)'
+	)
+	for cmd in "${!exp[@]}"
+	do
+		got=$(set +x; eval "$cmd" 2>&1)
+		[[ $got == "${exp[$cmd]}" ]] || err_exit "typeset -g in $(printf %q "$cmd") failed to activate global" \
+			"scope from name space (expected $(printf %q "${exp[$cmd]}"), got $(printf %q "$got"))"
+	done
+	unset exp
+fi
+
 # ======
 exit $((Errors<125?Errors:125))