26042 with some fixes from 26043 (Mikael):

allow <(...) and >(...) to occur in the middle of strings
and =(...) to have extra text following

7 files changed, 143 insertions, 61 deletions

diff --git a/ChangeLog b/ChangeLog
index 3912efe29..94f89b120 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,11 @@
+2008-11-13  Peter Stephenson  <p.w.stephenson@ntlworld.com>
+
+	* 26042 with some fixes from 26043 (Mikael): README,
+	Doc/Zsh/expn.yo, Src/exec.c, Src/lex.c, Src/subst.c,
+	Test/D03procsubst.zst: allow <(...) and >(...) to occur
+	in the middle of command arguments and =(...) to have
+	other strings following.
+
 2008-11-12  Oliver Kiddle  <opk@zsh.org>
 
 	* 26030: Src/Zle/zle_main.c: fix memory leak in vared
diff --git a/Doc/Zsh/expn.yo b/Doc/Zsh/expn.yo
index 5526ff3ea..2ab0d417e 100644
--- a/Doc/Zsh/expn.yo
+++ b/Doc/Zsh/expn.yo
@@ -353,12 +353,18 @@ texinode(Process Substitution)(Parameter Expansion)(History Expansion)(Expansion
 sect(Process Substitution)
 cindex(process substitution)
 cindex(substitution, process)
-Each command argument of the form
+Each part of a command argument that takes the form
 `tt(<LPAR())var(list)tt(RPAR())',
 `tt(>LPAR())var(list)tt(RPAR())' or
 `tt(=LPAR())var(list)tt(RPAR())'
-is subject to process substitution.
-In the case of the tt(<) or tt(>) forms, the shell runs process
+is subject to process substitution.  The expression may be preceeded
+or followed by other strings except that, to prevent clashes with
+commonly occurring strings and patterns, the last
+form must occur at the start of a command argument, and none of
+the forms may occur inside parentheses used for grouping of patterns or
+inside parameter substitutions.
+
+In the case of the tt(<) or tt(>) forms, the shell runs the commands in
 var(list) asynchronously.  If the system supports the tt(/dev/fd)
 mechanism, the command argument is the name of the device file
 corresponding to a file descriptor; otherwise, if the system supports named
diff --git a/README b/README
index b64000c26..3a8597033 100644
--- a/README
+++ b/README
@@ -69,6 +69,17 @@ always the right behaviour for the intended purpose of debugging and is
 consistent with recent versions of other shells.  The option
 DEBUG_BEFORE_CMD can be unset to revert to the previous behaviour.
 
+Previously, process substitutions of the form =(...), <(...) and >(...)
+were only handled if they appeared as separate command arguments.
+(However, the latter two forms caused the current argument to be
+terminated and a new one started even if they occurred in the middle of
+a string.)  Now all three may be followed by other strings, and the
+latter two may also be preceeded by other strings.  None may occur inside
+parameter substitutions, or inside parentheses used for grouping of
+patterns, in order to avoid clashes with cases where
+tt(<) or tt(>) where not treated specially in previous versions of the
+shell.
+
 In previous versions of the shell it was possible to use index 0 in an
 array or string subscript to refer to the same element as index 1 if the
 option KSH_ARRAYS was not in effect.  This was a limited approximation to
diff --git a/Src/exec.c b/Src/exec.c
index 47c0184aa..a398211d3 100644
--- a/Src/exec.c
+++ b/Src/exec.c
@@ -3560,7 +3560,7 @@ readoutput(int in, int qt)
 
 /**/
 static Eprog
-parsecmd(char *cmd)
+parsecmd(char *cmd, char **eptr)
 {
     char *str;
     Eprog prog;
@@ -3571,7 +3571,9 @@ parsecmd(char *cmd)
 	return NULL;
     }
     *str = '\0';
-    if (str[1] || !(prog = parse_string(cmd + 2, 0))) {
+    if (eptr)
+	*eptr = str+1;
+    if (!(prog = parse_string(cmd + 2, 0))) {
 	zerr("parse error in process substitution");
 	return NULL;
     }
@@ -3582,7 +3584,7 @@ parsecmd(char *cmd)
 
 /**/
 char *
-getoutputfile(char *cmd)
+getoutputfile(char *cmd, char **eptr)
 {
     pid_t pid;
     char *nam;
@@ -3592,7 +3594,7 @@ getoutputfile(char *cmd)
 
     if (thisjob == -1)
 	return NULL;
-    if (!(prog = parsecmd(cmd)))
+    if (!(prog = parsecmd(cmd, eptr)))
 	return NULL;
     if (!(nam = gettempname(NULL, 0)))
 	return NULL;
@@ -3677,7 +3679,7 @@ namedpipe(void)
 
 /**/
 char *
-getproc(char *cmd)
+getproc(char *cmd, char **eptr)
 {
 #if !defined(HAVE_FIFOS) && !defined(PATH_DEV_FD)
     zerr("doesn't look like your system supports FIFOs.");
@@ -3696,7 +3698,7 @@ getproc(char *cmd)
 	return NULL;
     if (!(pnam = namedpipe()))
 	return NULL;
-    if (!(prog = parsecmd(cmd)))
+    if (!(prog = parsecmd(cmd, eptr)))
 	return NULL;
     if (!jobtab[thisjob].filelist)
 	jobtab[thisjob].filelist = znewlinklist();
@@ -3723,7 +3725,7 @@ getproc(char *cmd)
     if (thisjob == -1)
 	return NULL;
     pnam = hcalloc(strlen(PATH_DEV_FD) + 6);
-    if (!(prog = parsecmd(cmd)))
+    if (!(prog = parsecmd(cmd, eptr)))
 	return NULL;
     mpipe(pipes);
     if ((pid = zfork(&bgtime))) {
@@ -3772,7 +3774,7 @@ getpipe(char *cmd, int nullexec)
     pid_t pid;
     struct timeval bgtime;
 
-    if (!(prog = parsecmd(cmd)))
+    if (!(prog = parsecmd(cmd, NULL)))
 	return -1;
     mpipe(pipes);
     if ((pid = zfork(&bgtime))) {
diff --git a/Src/lex.c b/Src/lex.c
index 025387ca1..f5999d798 100644
--- a/Src/lex.c
+++ b/Src/lex.c
@@ -835,7 +835,7 @@ gettok(void)
 	return OUTPAR;
     case LX1_INANG:
 	d = hgetc();
-	if (!incmdpos && d == '(') {
+	if (d == '(') {
 	    hungetc(d);
 	    lexstop = 0;
 	    unpeekfd:
@@ -1152,20 +1152,13 @@ gettokstr(int c, int sub)
 		c = Comma;
 	    break;
 	case LX2_OUTANG:
-	    if (!intpos) {
-		if (in_brace_param || sub)
-		    break;
-		else
-		    goto brk;
-	    }
+	    if (in_brace_param || sub)
+		break;
 	    e = hgetc();
 	    if (e != '(') {
 		hungetc(e);
 		lexstop = 0;
-		if (in_brace_param || sub)
-		    break;
-		else
-		    goto brk;
+		goto brk;
 	    }
 	    add(Outang);
 	    if (skipcomm()) {
@@ -1178,7 +1171,7 @@ gettokstr(int c, int sub)
 	    if (isset(SHGLOB) && sub)
 		break;
 	    e = hgetc();
-	    if(e == '(' && intpos) {
+	    if (!(in_brace_param || sub) && e == '(') {
 		add(Inang);
 		if (skipcomm()) {
 		    peek = LEXERR;
diff --git a/Src/subst.c b/Src/subst.c
index d76215838..5cc4748b8 100644
--- a/Src/subst.c
+++ b/Src/subst.c
@@ -56,43 +56,27 @@ prefork(LinkList list, int flags)
 
     queue_signals();
     for (node = firstnode(list); node; incnode(node)) {
-	char *str, c;
-
-	str = (char *)getdata(node);
-	if (((c = *str) == Inang || c == Outang || c == Equals) &&
-	    str[1] == Inpar) {
-	    if (c == Inang || c == Outang)
-		setdata(node, (void *) getproc(str));	/* <(...) or >(...) */
-	    else
-		setdata(node, (void *) getoutputfile(str));	/* =(...) */
-	    if (!getdata(node)) {
-		setdata(node, dupstring(""));
-		unqueue_signals();
-		return;
-	    }
-	} else {
-	    if (isset(SHFILEEXPANSION)) {
-		/*
-		 * Here and below we avoid taking the address
-		 * of a void * and then pretending it's a char **
-		 * instead of a void ** by a little inefficiency.
-		 * This could be avoided with some extra linked list
-		 * machinery, but that would need quite a lot of work
-		 * to ensure consistency.  What we really need is
-		 * templates...
-		 */
-		char *cptr = (char *)getdata(node);
-		filesub(&cptr, flags & (PF_TYPESET|PF_ASSIGN));
-		/*
-		 * The assignment is so simple it's not worth
-		 * testing if cptr changed...
-		 */
-		setdata(node, cptr);
-	    }
-	    if (!(node = stringsubst(list, node, flags & PF_SINGLE, asssub))) {
-		unqueue_signals();
-		return;
-	    }
+	if (isset(SHFILEEXPANSION)) {
+	    /*
+	     * Here and below we avoid taking the address
+	     * of a void * and then pretending it's a char **
+	     * instead of a void ** by a little inefficiency.
+	     * This could be avoided with some extra linked list
+	     * machinery, but that would need quite a lot of work
+	     * to ensure consistency.  What we really need is
+	     * templates...
+	     */
+	    char *cptr = (char *)getdata(node);
+	    filesub(&cptr, flags & (PF_TYPESET|PF_ASSIGN));
+	    /*
+	     * The assignment is so simple it's not worth
+	     * testing if cptr changed...
+	     */
+	    setdata(node, cptr);
+	}
+	if (!(node = stringsubst(list, node, flags & PF_SINGLE, asssub))) {
+	    unqueue_signals();
+	    return;
 	}
     }
     for (node = firstnode(list); node; incnode(node)) {
@@ -168,7 +152,37 @@ stringsubst(LinkList list, LinkNode node, int ssub, int asssub)
     char *str  = str3, c;
 
     while (!errflag && (c = *str)) {
-	if ((qt = c == Qstring) || c == String) {
+	if ((c == Inang || c == Outang || (str == str3 && c == Equals)) &&
+	    str[1] == Inpar) {
+	    char *subst, *rest, *snew, *sptr;
+	    int str3len = str - str3, sublen, restlen;
+
+	    if (c == Inang || c == Outang)
+		subst = getproc(str, &rest);	/* <(...) or >(...) */
+	    else
+		subst = getoutputfile(str, &rest);	/* =(...) */
+	    if (!subst)
+		subst = "";
+
+	    sublen = strlen(subst);
+	    restlen = strlen(rest);
+	    sptr = snew = hcalloc(str3len + sublen + restlen + 1);
+	    if (str3len) {
+		memcpy(sptr, str3, str3len);
+		sptr += str3len;
+	    }
+	    if (sublen) {
+		memcpy(sptr, subst, sublen);
+		sptr += sublen;
+	    }
+	    if (restlen)
+		memcpy(sptr, rest, restlen);
+	    sptr[restlen] = '\0';
+	    str3 = snew;
+	    str = snew + str3len + sublen;
+	    setdata(node, str3);
+	    continue;
+	} else if ((qt = c == Qstring) || c == String) {
 	    if ((c = str[1]) == Inpar) {
 		if (!qt)
 		    list->list.flags |= LF_ARRAY;
diff --git a/Test/D03procsubst.ztst b/Test/D03procsubst.ztst
index e176d8934..37a67630f 100644
--- a/Test/D03procsubst.ztst
+++ b/Test/D03procsubst.ztst
@@ -36,3 +36,51 @@
 0:FDs remain open for external commands called from functions
 >First
 >Zweite
+
+  catfield2() {
+    local -a args
+    args=(${(s.,.)1})
+    print $args[1]
+    cat $args[2]
+    print $args[3]
+  }
+  catfield2 up,<(print $'\x64'own),sideways
+0:<(...) when embedded within an argument
+>up
+>down
+>sideways
+
+  outputfield2() {
+    local -a args
+    args=(${(s.,.)1})
+    print $args[1]
+    echo 'How sweet the moonlight sits upon the bank' >$args[2]
+    print $args[3]
+  }
+  outputfield2 muddy,>(sed -e s/s/th/g >outputfield2.txt),vesture
+  # yuk
+  while [[ ! -e outputfield2.txt || ! -s outputfield2.txt ]]; do :; done
+  cat outputfield2.txt
+0:>(...) when embedded within an argument
+>muddy
+>vesture
+>How thweet the moonlight thitth upon the bank
+
+  catfield1() {
+    local -a args
+    args=(${(s.,.)1})
+    cat $args[1]
+    print $args[2]
+  }
+  catfield1 =(echo s$'\x69't),jessica
+0:=(...) followed by something else without a break
+>sit
+>jessica
+
+  (
+  setopt nonomatch
+  # er... why is this treated as a glob?
+  print everything,=(here is left),alone
+  )
+0:=(...) preceded by other stuff has no special effect
+>everything,=(here is left),alone