Backport -P and -t flags for 'type'/'whence' from ksh93v- (#392)

This commit backports the whence '-t' option from ksh93v-. The '-t'
option is useful when one needs to identify the type of a command.
The '-t' flag was added by ksh93v- for compatibility with Bash.

It should be noted the ksh93v- patch had one bug, which this commit
fixes. Path-bound builtins from /opt/ast/bin were classified as
files if loaded from /opt/ast/bin in the PATH. Reproducer:
   $ PATH=/opt/ast/bin whence -t cat
   file

src/cmd/ksh93/bltins/whence.c:
- Simplify the bitmask values for the command and whence builtin
  flags.
- Add the -t flag to the whence and type builtins. To prevent bugs,
  -t will always override -v if both of those flags were passed.

src/cmd/ksh93/data/builtins.c,
src/cmd/ksh93/sh.1:
- Add documentation for the new -t option.

ANNOUNCE

@@ -35,6 +35,18 @@ To get involved in development, read the brief policy information in
 README.md and then jump right in with a pull request or email a patch.
 See the TODO file in the top-level directory for a to-do list.
 
+### MAIN CHANGES between 1.0.0-beta.2 and 1.0.0-(unreleased) ###
+
+New features in built-in commands:
+
+- Two bash-like flags for 'whence' were backported from ksh 93v-:
+  - 'whence -P/type -P' is an alias to the existing -p flag.
+  - 'whence -t/type -t' will print only the type of a command in a simple
+    format that is designed to be easy to use for scripts. Example:
+        $ type -t typeset; whence -t sh
+        builtin
+        file
+
 ### MAIN CHANGES between 1.0.0-beta.1 and 1.0.0-beta.2 ###
 
 New features in built-in commands:

NEWS

@@ -5,6 +5,14 @@ Any uppercase BUG_* names are modernish shell bug IDs.
 
 2021-12-27:
 
+- Two bash-like flags for 'whence' were backported from ksh 93v-:
+  - 'whence -P/type -P' is an alias to the existing -p flag.
+  - 'whence -t/type -t' will print only the type of a command in a simple
+    format that is designed to be easy to use for scripts. Example:
+        $ type -t typeset; whence -t sh
+        builtin
+        file
+
 - Fixed a crash or freeze that would occur on Linux when using Ctrl+C to
   interrupt a command substitution containing a pipe in an interactive shell.
 

src/cmd/ksh93/bltins/whence.c

@@ -20,7 +20,8 @@
 ***********************************************************************/
 /*
  * command [-pvVx] name [arg...]
- * whence [-afpqv] name...
+ * type [-afpPqt] name...
+ * whence [-afpPqtv] name...
  *
  *   David Korn
  *   AT&T Labs
@@ -35,12 +36,13 @@
 #include	"shlex.h"
 #include	"builtins.h"
 
-#define P_FLAG	1
+#define P_FLAG	(1 << 0)
-#define V_FLAG	2
+#define V_FLAG	(1 << 1)
-#define A_FLAG	4
+#define A_FLAG	(1 << 2)
-#define F_FLAG	010
+#define F_FLAG	(1 << 3)
-#define X_FLAG	020
+#define X_FLAG	(1 << 4)
-#define Q_FLAG	040
+#define Q_FLAG	(1 << 5)
+#define T_FLAG	(1 << 6)
 
 static int whence(Shell_t *,char**, int);
 
@@ -121,12 +123,15 @@ int	b_whence(int argc,char *argv[],Shbltin_t *context)
 	    case 'v':
 		flags |= V_FLAG;
 		break;
+	    case 't':
+		flags |= T_FLAG;
+		break;
 	    case 'f':
 		flags |= F_FLAG;
 		break;
+	    case 'P':
 	    case 'p':
 		flags |= P_FLAG;
-		flags &= ~V_FLAG;
 		break;
 	    case 'q':
 		flags |= Q_FLAG;
@@ -138,6 +143,8 @@ int	b_whence(int argc,char *argv[],Shbltin_t *context)
 		errormsg(SH_DICT,ERROR_usage(2), "%s", opt_info.arg);
 		UNREACHABLE();
 	}
+	if(flags&(P_FLAG|T_FLAG))
+		flags &= ~V_FLAG;
 	argv += opt_info.index;
 	if(error_info.errors || !*argv)
 	{
@@ -171,7 +178,10 @@ static int whence(Shell_t *shp,char **argv, register int flags)
 		/* reserved words first */
 		if(sh_lookup(name,shtab_reserved))
 		{
-			sfprintf(sfstdout,"%s%s\n",name,(flags&V_FLAG)?sh_translate(is_reserved):"");
+			if(flags&T_FLAG)
+				sfprintf(sfstdout,"keyword\n");
+			else
+				sfprintf(sfstdout,"%s%s\n",name,(flags&V_FLAG)?sh_translate(is_reserved):"");
 			if(!aflag)
 				continue;
 			aflag++;
@@ -186,7 +196,10 @@ static int whence(Shell_t *shp,char **argv, register int flags)
 				msg = sh_translate(is_alias);
 				sfprintf(sfstdout,msg,name);
 			}
-			sfputr(sfstdout,sh_fmtq(cp),'\n');
+			if(flags&T_FLAG)
+				sfputr(sfstdout,"alias",'\n');
+			else
+				sfputr(sfstdout,sh_fmtq(cp),'\n');
 			if(!aflag)
 				continue;
 			cp = 0;
@@ -198,7 +211,8 @@ static int whence(Shell_t *shp,char **argv, register int flags)
 		{
 			if(flags&Q_FLAG)
 				continue;
-			sfputr(sfstdout,name,-1);
+			if(!(flags&T_FLAG))
+				sfputr(sfstdout,name,-1);
 			if(flags&V_FLAG)
 			{
 				if(nv_isnull(np))
@@ -213,6 +227,8 @@ static int whence(Shell_t *shp,char **argv, register int flags)
 				else
 					sfprintf(sfstdout,sh_translate(is_function));
 			}
+			else if(flags&T_FLAG)
+				sfprintf(sfstdout,"function");
 			sfputc(sfstdout,'\n');
 			if(!aflag)
 				continue;
@@ -230,7 +246,10 @@ static int whence(Shell_t *shp,char **argv, register int flags)
 				cp = "";
 			if(flags&Q_FLAG)
 				continue;
-			sfprintf(sfstdout,"%s%s\n",name,cp);
+			if(flags&T_FLAG)
+				sfprintf(sfstdout,"builtin\n");
+			else
+				sfprintf(sfstdout,"%s%s\n",name,cp);
 			if(!aflag)
 				continue;
 			aflag++;
@@ -273,7 +292,9 @@ static int whence(Shell_t *shp,char **argv, register int flags)
 				{
 					/* Undefined/autoloadable function on FPATH */
 					sfputr(sfstdout,sh_fmtq(cp),-1);
-					if(flags&V_FLAG)
+					if(flags&T_FLAG)
+						sfprintf(sfstdout,"function");
+					else if(flags&V_FLAG)
 					{
 						sfprintf(sfstdout,sh_translate(is_ufunction));
 						sfprintf(sfstdout,sh_translate(e_autoloadfrom),sh_fmtq(stakptr(PATH_OFFSET)));
@@ -283,13 +304,20 @@ static int whence(Shell_t *shp,char **argv, register int flags)
 			}
 			else if(cp)
 			{
+				int is_pathbound_builtin = 0;
 				cp = path_fullname(shp,cp);  /* resolve '.' & '..' */
-				if(flags&V_FLAG)
+				if(flags&(V_FLAG|T_FLAG))
 				{
-					sfputr(sfstdout,sh_fmtq(name),' ');
+					if(!(flags&T_FLAG))
+						sfputr(sfstdout,sh_fmtq(name),' ');
 					/* built-in version of program */
 					if(nv_search(cp,shp->bltin_tree,0))
-						msg = sh_translate(is_builtver);
+					{
+						if(flags&T_FLAG)
+							is_pathbound_builtin = 1;
+						else
+							msg = sh_translate(is_builtver);
+					}
 					/* tracked aliases next */
 					else if(!sh_isstate(SH_DEFPATH)
 					&& (np = nv_search(name,shp->track_tree,0))
@@ -298,9 +326,13 @@ static int whence(Shell_t *shp,char **argv, register int flags)
 						msg = sh_translate(is_talias);
 					else
 						msg = sh_translate("is");
-					sfputr(sfstdout,msg,' ');
+					if(!(flags&T_FLAG))
+						sfputr(sfstdout,msg,' ');
 				}
-				sfputr(sfstdout,sh_fmtq(cp),'\n');
+				if(flags&T_FLAG)
+					sfputr(sfstdout,is_pathbound_builtin ? "builtin" : "file",'\n');
+				else
+					sfputr(sfstdout,sh_fmtq(cp),'\n');
 				free((char*)cp);
 			}
 			else if(aflag<=1) 

src/cmd/ksh93/data/builtins.c

@@ -2051,7 +2051,7 @@ _JOB_
 ;
 
 const char sh_optwhence[] =
-"[-1c?\n@(#)$Id: whence (ksh 93u+m) 2020-09-25 $\n]"
+"[-1c?\n@(#)$Id: whence (ksh 93u+m) 2021-12-27 $\n]"
 "[--catalog?" SH_DICT "]"
 "[+NAME?whence, type - locate a command and describe its type]"
 "[+DESCRIPTION?Without \b-v\b, \bwhence\b writes on standard output an "
@@ -2064,8 +2064,11 @@ const char sh_optwhence[] =
 "[+?The \btype\b command is equivalent to \bwhence -v\b.]"
 "[a?Like \b-v\b but displays all uses for each \aname\a rather than the first.]"
 "[f?Do not check for functions.]"
-"[p?Do not check to see if \aname\a is a reserved word, a built-in, "
+"[p|P?Do not check to see if \aname\a is a reserved word, a built-in, "
 	"an alias, or a function.  This turns off the \b-v\b option.]"
+"[t?Output only a single-word type indicator for each \aname\a found: "
+	"\bkeyword\b, \balias\b, \bbuiltin\b, \bfunction\b or \bfile\b. "
+	"This turns off the \b-v\b option.]"
 "[q?Quiet mode. Returns 0 if all arguments are built-ins, functions, or are "
 	"programs found on the path.]"
 "[v?For each name you specify, the shell displays a line that indicates "

src/cmd/ksh93/data/testops.c

@@ -103,11 +103,7 @@ const char sh_opttest[] =
 	"[+-e \afile\a?\afile\a exists and is not a broken symlink.]"
 	"[+-f \afile\a?\afile\a is a regular file.]"
 	"[+-g \afile\a?\afile\a has its set-group-id bit set.]"
-	"[+-h \afile\a?Same as \b-L\b.]"
 	"[+-k \afile\a *?\afile\a has its sticky bit on.]"
-#if SHOPT_TEST_L
-	"[+-l \afile\a *?Same as \b-L\b.]"
-#endif
 	"[+-n \astring\a?Length of \astring\a is non-zero.]"
 	"[+-o \aoption\a *?Shell option \aoption\a is enabled.]"
 	"[+-p \afile\a?\afile\a is a FIFO.]"
@@ -122,7 +118,11 @@ const char sh_opttest[] =
 	"[+-z \astring\a?\astring\a is a zero-length string.]"
 	"[+-G \afile\a *?Group of \afile\a is the effective "
 		"group ID of the current process.]"
-	"[+-L \afile\a?\afile\a is a symbolic link.]"
+#if SHOPT_TEST_L
+	"[+-L|l|h \afile\a?\afile\a is a symbolic link.]"
+#else
+	"[+-L|h \afile\a?\afile\a is a symbolic link.]"
+#endif
 	"[+-N \afile\a *?\afile\a has been modified since it was last read.]"
 	"[+-O \afile\a *?\afile\a exists and owner is the effective "
 		"user ID of the current process.]"

src/cmd/ksh93/sh.1

@@ -7672,7 +7672,7 @@ Does nothing, and exits 0. Used with
 .B while
 for infinite loops.
 .TP
-\f3type\fP \*(OK \f3\-afpq\fP \*(CK \f2name\^\fP .\|.\|.
+\f3type\fP \*(OK \f3\-afpPqt\fP \*(CK \f2name\^\fP .\|.\|.
 The same as
 .BR whence\ \-v .
 .TP
@@ -8292,7 +8292,7 @@ See
 for a description of the format of
 .IR job .
 .TP
-\f3whence\fP \*(OK \f3\-afpqv\fP \*(CK \f2name\^\fP .\|.\|.
+\f3whence\fP \*(OK \f3\-afpPqtv\fP \*(CK \f2name\^\fP .\|.\|.
 For each
 .IR name ,
 indicate how it
@@ -8307,13 +8307,13 @@ The
 option skips the search for functions.
 The
 .B \-p
-option
+and
-does a path search for
+.B \-P
+options
+do a path search for
 .I name\^
 even if name is an alias, a function, or a reserved word.
-The
+Both of these options turn off the
-.B \-p
-option turns off the
 .B \-v
 option.
 The
@@ -8325,6 +8325,17 @@ to enter quiet mode.
 will return zero if all arguments are built-ins, functions, or
 are programs found on the path.
 The
+.B \-t
+option only outputs the type of the given command.
+Like
+.B \-p
+and
+.BR \-P ,
+.B \-t
+will turn off the
+.B \-v
+option.
+The
 .B \-a
 option
 is similar to the

src/cmd/ksh93/tests/builtins.sh

@@ -838,6 +838,89 @@ whence -q cat nonexist && err_exit "'whence -q' has the wrong exit status"
 whence -q nonexist && err_exit "'whence -q' has the wrong exit status"
 PATH=$save_PATH
 
+# ======
+# These are the regression tests for the whence builtin's '-t' flag
+for w in 'whence -t' 'type -t' 'whence -t -v'; do
+	exp=file
+	got=$($w $SHELL)
+	[[ $exp == "$got" ]] || err_exit "'$w' has the wrong output for external commands" \
+		"(expected $(printf %q "$exp"); got $(printf %q "$got"))"
+	got=$(builtin -d chmod; hash chmod; $w chmod)
+	[[ $exp == "$got" ]] || err_exit "'$w' has the wrong output for tracked aliases" \
+		"(expected $(printf %q "$exp"); got $(printf %q "$got"))"
+	exp=keyword
+	got=$($w time)
+	[[ $exp == "$got" ]] || err_exit "'$w' has the wrong output for keywords" \
+		"(expected $(printf %q "$exp"); got $(printf %q "$got"))"
+	exp=builtin
+	got=$($w sleep)
+	[[ $exp == "$got" ]] || err_exit "'$w' has the wrong output for regular builtins" \
+		"(expected $(printf %q "$exp"); got $(printf %q "$got"))"
+	got=$($w export)
+	[[ $exp == "$got" ]] || err_exit "'$w' has the wrong output for special builtins" \
+		"(expected $(printf %q "$exp"); got $(printf %q "$got"))"
+	if [[ $(PATH=/opt/ast/bin type cat) != "cat is a shell builtin version of /opt/ast/bin/cat" ]]
+	then	warning "/opt/ast/bin/cat isn't a builtin; skipping path-bound builtin tests"
+	else
+		got=$(PATH=/opt/ast/bin $w cat)
+		[[ $exp == "$got" ]] || err_exit "Test A: '$w' has the wrong output for path-bound builtins" \
+			"(expected $(printf %q "$exp"); got $(printf %q "$got"))"
+		got=$($w /opt/ast/bin/cat)
+		[[ $exp == "$got" ]] || err_exit "Test B: '$w' has the wrong output for path-bound builtins" \
+			"(expected $(printf %q "$exp"); got $(printf %q "$got"))"
+		got=$(PATH=/opt/ast/bin:$PATH; $w cat)
+		[[ $exp == "$got" ]] || err_exit "Test C: '$w' has the wrong output for path-bound builtins" \
+			"(expected $(printf %q "$exp"); got $(printf %q "$got"))"
+		got=$(builtin cat; $w cat)
+		[[ $exp == "$got" ]] || err_exit "Test D: '$w' has the wrong output for path-bound builtins" \
+			"(expected $(printf %q "$exp"); got $(printf %q "$got"))"
+	fi
+	exp=alias
+	got=$(alias foo=bar; $w foo)
+	[[ $exp == "$got" ]] || err_exit "'$w' has the wrong output for aliases" \
+		"(expected $(printf %q "$exp"); got $(printf %q "$got"))"
+	exp=function
+	got=$(foo() { true; }; $w foo)
+	[[ $exp == "$got" ]] || err_exit "'$w' has the wrong output for POSIX functions" \
+		"(expected $(printf %q "$exp"); got $(printf %q "$got"))"
+	got=$(function foo { true; }; $w foo)
+	[[ $exp == "$got" ]] || err_exit "'$w' has the wrong output for KornShell functions" \
+		"(expected $(printf %q "$exp"); got $(printf %q "$got"))"
+	got=$(autoload FooBar; $w FooBar)
+	[[ $exp == "$got" ]] || err_exit "'$w' has the wrong output for undefined autoloaded functions" \
+		"(expected $(printf %q "$exp"); got $(printf %q "$got"))"
+	echo 'FooBar() { true; }' > FooBar
+	got=$(FPATH=. autoload FooBar; $w FooBar)
+	[[ $exp == "$got" ]] || err_exit "'$w' has the wrong output for autoloaded POSIX functions" \
+		"(expected $(printf %q "$exp"); got $(printf %q "$got"))"
+	echo 'function FooBar { true; }' > FooBar
+	got=$(FPATH=. autoload FooBar; $w FooBar)
+	[[ $exp == "$got" ]] || err_exit "'$w' has the wrong output for autoloaded KornShell functions" \
+		"(expected $(printf %q "$exp"); got $(printf %q "$got"))"
+
+	# The final few tests are for '-t -a'
+	exp="alias
+function
+builtin
+$($w -pa cat)"
+	got=$(alias cat=false
+		autoload cat
+		PATH=/opt/ast/bin:$PATH $w -a cat)
+	[[ $exp == "$got" ]] || err_exit "'$w -a' output is incorrect (cat command)" \
+		"(expected $(printf %q "$exp"); got $(printf %q "$got"))"
+	if [[ -n $($w -pa time) ]]
+	then	exp="keyword
+alias
+$($w -pa time)"
+	else
+		exp=$'keyword\nalias'
+	fi
+	got=$(alias time=nottime
+		$w -a time)
+	[[ $exp == "$got" ]] || err_exit "'$w -a' output is incorrect (time command)" \
+		"(expected $(printf %q "$exp"); got $(printf %q "$got"))"
+done
+
 # ======
 # 'cd ../.foo' should not exclude the '.' in '.foo'
 # https://bugzilla.redhat.com/889748