From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <zsh-workers-return-9154-mason-zsh=primenet.com.au@sunsite.auc.dk>
Received: (qmail 6494 invoked from network); 25 Dec 1999 00:41:18 -0000
Received: from sunsite.auc.dk (130.225.51.30)
  by ns1.primenet.com.au with SMTP; 25 Dec 1999 00:41:18 -0000
Received: (qmail 23611 invoked by alias); 25 Dec 1999 00:41:13 -0000
Mailing-List: contact zsh-workers-help@sunsite.auc.dk; run by ezmlm
Precedence: bulk
X-No-Archive: yes
X-Seq: 9154
Received: (qmail 23604 invoked from network); 25 Dec 1999 00:41:11 -0000
To: zsh-workers@sunsite.auc.dk
Subject: PATCH: regexparse
MIME-Version: 1.0 (generated by AKEMI 1.13.2 - =?ISO-2022-JP?B?Ig==?=
 =?ISO-2022-JP?B?GyRCQTA0Y0s8GyhCIg==?=)
Content-Type: text/plain; charset=US-ASCII
From: Tanaka Akira <akr@jaist.ac.jp>
Date: 25 Dec 1999 09:41:08 +0900
Message-ID: <rsqiu1nj1i3.fsf@crane.jaist.ac.jp>
User-Agent: Chao-gnus/6.12.5 AKEMI/1.13.2 (=?ISO-2022-JP?B?GyRCQTAbKEI=?=
 =?ISO-2022-JP?B?GyRCNGNLPBsoQg==?=) FLAM-DOODLE/1.12.6
 (=?ISO-2022-JP?B?GyRCM3cbKEI=?= 10R4.0/5.0) Emacs/20.4
 (sparc-sun-solaris2.6) MULE/4.0 (HANANOEN)

Finally, I re-implemented most of `_regex_arguments' in C.
This should accelerate `_apt'.

This patch adds a builtin `regexparse' to the module `zsh/zutil' for
internal use of `_regex_arguments'.

I select `zsh/zutil' because `regexparse' doesn't depend on completion
codes --- it just executes strings supplied by arguments which may
have compadd, etc.  Its own module is another option, though.

Index: Completion/Base/_regex_arguments
===================================================================
RCS file: /projects/zsh/zsh/Completion/Base/_regex_arguments,v
retrieving revision 1.1.1.10
diff -u -r1.1.1.10 _regex_arguments
--- Completion/Base/_regex_arguments	1999/12/10 14:47:55	1.1.1.10
+++ Completion/Base/_regex_arguments	1999/12/25 00:23:52
@@ -22,16 +22,17 @@
 # pattern = "/" ( glob | "[]" ) "/" [ "+" | "-" ]
 # lookahead = "%" glob "%"
 # guard = "-" zsh-code-to-eval
-# action = ":" zsh-code-to-eval
+# caction = ":" zsh-code-to-eval
+# action = "{" zsh-code-to-eval "}"
 
 ## regex word sequence definition:
 
-# element = pattern [ lookahead ] [ guard ] [ action ]
-#
+# element = pattern [ lookahead ] [ guard ] [ caction ]
+# 
 # regex = element 
 #	| "(" regex ")"
 #	| regex "#"
-#	| regex regex
+#	| ( regex | action ) #
 #	| regex "|" regex
 
 # example:
@@ -56,323 +57,38 @@
 # 1 : parse error
 # 2 : fatal parse error
 
-_ra_parse_elt () {
-  local state act
-  if (( $#regex < index )); then
-    return 1
-  else
-    case "$regex[index]" in
-      /*/([-+]|)) state=$index
-          first=($state)
-	  last=($state)
-	  nullable=
-	  case "$regex[index]" in 
-	    */+) cutoff[$state]=+;;
-	    */) cutoff[$state]=/;;
-	    */-) cutoff[$state]=-;;
-	  esac
-          pattern[$state]="${${regex[index++]#/}%/([-+]|)}"
-	  if [[ $pattern[$state] != "[]" ]]; then
-	    pattern[$state]="(#b)((#B)$pattern[$state])"
-	  fi
-	  if [[ $index -le $#regex && $regex[index] = %*% ]]; then
-	    lookahead[$state]="(#B)${regex[index++][2,-2]}"
-	  else
-	    lookahead[$state]=""
-	  fi
-	  if [[ $index -le $#regex && $regex[index] = -* ]]; then
-	    guard[$state]="${regex[index++][2,-1]}"
-	  else
-	    guard[$state]=""
-	  fi
-	  if [[ $index -le $#regex && $regex[index] = :* ]]; then
-	    act="${regex[index++][2,-1]}"
-	    action[$state]="$act"
-	    : ${actions[$act]::="${actions[$act]} $state"}
-	  else
-	    action[$state]=""
-	  fi
-	  ;;
-      \() (( index++ ))
-          _ra_parse_alt || return $?
-	  [[ $index -le $#regex && "$regex[$index]" = \) ]] || return 2
-	  (( index++ ))
-	  ;;
-      *)  return 1
-          ;;
-    esac
-  fi
-
-  return 0
-}
-
-_ra_parse_clo () {
-  _ra_parse_elt || return $?
-
-  if (( index <= $#regex )) && [[ "$regex[$index]" = \# ]]; then
-    (( index++ ))
-    nullable=yes
-
-    for i in $last; do tbl[$i]="$tbl[$i] $first"; done
-  fi
-
-  return 0
+_ra_comp () {
+  _ra_actions=("$_ra_actions[@]" "$1")
 }
 
-_ra_parse_seq () {
-  local last_seq
-  local first_seq nullable_seq
-  first_seq=()
-  nullable_seq=yes
-
-  _ra_parse_clo || {
-    if (( $? == 2 )); then
-      return 2
-    else
-      first=()
-      last=()
-      nullable=yes
-      return 0
-    fi
-  }
-  first_seq=($first)
-  last_seq=($last)
-  [[ -n "$nullable" ]] || nullable_seq=
-
-  while :; do
-    _ra_parse_clo || {
-      if (( $? == 2 )); then
-        return 2
-      else
-        break
-      fi
-    }
-    for i in $last_seq; do tbl[$i]="${tbl[$i]} $first"; done
-    [[ -n "$nullable_seq" ]] && first_seq=($first_seq $first)
-    [[ -n "$nullable" ]] || { nullable_seq= last_seq=() }
-    last_seq=($last_seq $last)
-  done
-
-  first=($first_seq)
-  nullable=$nullable_seq
-  last=($last_seq)
-  return 0
-}
-
-_ra_parse_alt () {
-  local last_alt
-  local first_alt nullable_alt 
-  first_alt=()
-  nullable_alt=
-
-  _ra_parse_seq || return $?
-  first_alt=($first_alt $first)
-  last_alt=($last_alt $last)
-  [[ -n "$nullable" ]] && nullable_alt=yes
-
-  while :; do
-    (( index <= $#regex )) || break
-    [[ "$regex[$index]" = \| ]] || break
-    (( index++ ))
-
-    _ra_parse_seq || {
-      if (( $? == 2 )); then
-        return 2
-      else
-        break
-      fi
-    }
-    first_alt=($first_alt $first)
-    last_alt=($last_alt $last)
-    [[ -n "$nullable" ]] && nullable_alt=yes
-  done
-
-  first=($first_alt)
-  last=($last_alt)
-  nullable=$nullable_alt
-  return 0
-}
-
-## function generator
-
-_ra_gen_func () {
-  local old new
-  local state index
-  local test tmp
-  local start="0"
-
-  old=()
-  new=($start)
-
-  print -lr - \
-    "$funcname () {" \
-      'local _ra_state _ra_left _ra_right _ra_actions' \
-      "_ra_state=$start" \
-      '_ra_left=' \
-      '_ra_right="${(pj:\0:)${(@)words[1,CURRENT - 1]:Q}}"$'\''\0'\''"$PREFIX"' \
-      '_ra_actions=()' \
-      'while :; do' \
-	'case "$_ra_state" in'
-
-  while (( $#new )); do
-    state="$new[1]"
-    shift new
-    old=("$old[@]" "$state")
-    print -lr - \
-	"$state)"
-    _ra_gen_parse_state
-    print -lr - \
-	  ';;'
-  done
-
-  print -lr - \
-	'esac' \
-      'done' \
-      'while (( $#_ra_actions )); do' \
-	'case "$_ra_actions[1]" in'
-
-  for tmp in "${(@k)actions}"; do
-    if [[ "$tmp" != '' ]]; then
-      print -lr - "${(j:);&:)${=actions[$tmp]}})" $tmp ';;'
-    fi
-  done
-
-  print -lr - \
-	'esac' \
-	'shift _ra_actions' \
-      'done' \
-    '}'
-}
-
-_ra_gen_parse_state () {
-  local actions i p
-  test='if'
-  for index in $=tbl[$state]; do
-    if [[ "$pattern[$index]" != "[]" ]]; then
-      p="$pattern[$index]$lookahead[$index]*"
-      if [[ -z "$guard[$index]" ]]; then
-	print -lr - \
-	  "$test [[ \$_ra_right = \${~:-${(qqqq)p}} ]]"
-      else
-	print -lr - \
-	  "$test [[ \$_ra_right = \${~:-${(qqqq)p}} ]] && {" \
-	      "$guard[$index]" \
-	    "}"
-      fi
-      test='elif'
-      (( $old[(I)$index] || $new[(I)$index] )) || new=($index "$new[@]")
-      print -lr - \
-	  "then" \
-	    "_ra_state=$index" \
-	    '_ra_right="${_ra_right[mend[1] + 1, -1]}"'
-      actions=()
-      for i in $=tbl[$index]; do
-	if [[ -n $action[$i] ]]; then
-	  actions=($actions $i)
-	fi
-      done
-      case "$cutoff[$index]" in
-	+) print -lr - \
-	    '_ra_left="$_ra_left$match[1]"'
-	  if (( $#actions )); then
-	    print -lr - \
-	    "_ra_actions=($actions \$_ra_actions)"
-	  fi
-          ;;
-	/) print -lr - \
-	    '_ra_left='
-	  print -lr - \
-	    'if (( mend[1] )); then' \
-	      "_ra_actions=($actions)"
-	  if (( $#actions )); then
-	    print -lr - \
-	    'else' \
-	      "_ra_actions=($actions \$_ra_actions)"
-	  fi
-	  print -lr - \
-	    'fi'
-	  ;;
-	-) print -lr - \
-	    '_ra_left=' \
-	    "_ra_actions=($actions)"
-	  ;;
-      esac
-    fi
-  done
-
-  if [[ $test != 'if' ]]; then
-  # Some branchs are exists. But all of them are failed.
-    print -lr - \
-	  'else' \
-	    'if [[ "$_ra_left$_ra_right" = *$'\''\0'\''* ]]; then' \
-	      '_message "parse failed before current word"' \
-	      '_ra_actions=()' \
-	    'else' \
-	      'compset -p $(( $#PREFIX - $#_ra_right - $#_ra_left ))' \
-	    'fi' \
-	    'break' \
-	  'fi'
-  else
-  # There are no branch.
-    print -lr - \
-	  '_message "no more arguments"' \
-	  '_ra_actions=()' \
-	  'break'
-  fi
-}
-
 _regex_arguments () {
-  local funcname="_regex_arguments_tmp"
-  local funcdef
-
-  typeset -A tbl cutoff pattern lookahead guard action actions
-  local regex index first last nullable
-  local i state next
-
-  local cache_dir
-  zstyle -s ":completion${curcontext}:regex" cache-path cache_dir
-  [[ -z "$cache_dir" ]] && cache_dir="$HOME/.zsh/regex_arguments"
-  local cache_file="$cache_dir/$1"
-  local cache_test
-
-  if ! [[ -f "$cache_file" ]] || ! source "$cache_file" "$@"; then
-    funcname="$1"
-
-    regex=("${(@)argv[2,-1]}")
-    index=1
-    tbl=()
-    pattern=()
-    lookahead=()
-    guard=()
-    action=()
-    actions=()
-    _ra_parse_alt
-
-    if (( $? == 2 || index != $#regex + 1 )); then
-      if (( index != $#regex + 1 )); then
-	print "regex parse error at $index: $regex[index]" >&2
+  local regex funcname="$1"
+  shift
+  regex=(${@/(#b):(*)/":_ra_comp ${(qqqq)match[1]}"})
+
+  eval \
+  "$funcname"' () {
+    local _ra_p1 _ra_p2 _ra_left _ra_right _ra_com
+    local _ra_actions _ra_line="${(pj:\0:)${(@)words[1,CURRENT - 1]:Q}}"$'\''\0'\''"$PREFIX"
+    regexparse _ra_p1 _ra_p2 "$_ra_line" '"${(j: :)${(qqqq)regex[@]}}"'
+    case "$?" in
+    0|2) _message "no more arguments";;
+    1) 
+      if [[ "$_ra_line[_ra_p1 + 1, -1]" = *$'\''\0'\''* ]]; then
+	_message "parse failed before current word"
       else
-	print "regex parse error at $index (end)" >&2
+	: $#PREFIX - $#_ra_line + $_ra_p1 : $_ra_p2
+	_ra_left="$_ra_line[_ra_p1 + 1, _ra_p2]"
+	_ra_right="$_ra_line[_ra_p2 + 1, -1]"
+	compset -p $(( $#PREFIX - $#_ra_line + $_ra_p1 ))
+	for _ra_com in "$_ra_actions[@]"; do
+	  eval "$_ra_com"
+	done
       fi
-      return 1
-    fi
-
-    tbl[0]=" $first"
-
-    unfunction "$funcname" 2>/dev/null
-    funcdef="$(_ra_gen_func)"
-
-    if [[ -d "$cache_dir" && -w "$cache_dir" ]]; then
-      print -lr - \
-	'if [[ $# -eq '$#' && "$*" = '"${(qqqq)*}"' ]]; then' \
-	"$funcdef" \
-	'true; else false; fi' > "${cache_file}.$HOST.$$"
-      source "${cache_file}.$HOST.$$" "$@"
-      mv "${cache_file}.$HOST.$$" "${cache_file}"
-    else
-      source =(print -lr - "$funcdef")
-    fi
-  fi
+      ;;
+    3) _message "invalid regex";;
+    esac
+  }'
 }
 
 _regex_arguments "$@"
Index: Src/Modules/zutil.c
===================================================================
RCS file: /projects/zsh/zsh/Src/Modules/zutil.c,v
retrieving revision 1.1.1.4
diff -u -r1.1.1.4 zutil.c
--- Src/Modules/zutil.c	1999/12/16 14:26:37	1.1.1.4
+++ Src/Modules/zutil.c	1999/12/25 00:23:55
@@ -721,9 +721,438 @@
     return 1;
 }
 
+/* Regexparse stuff. */
+
+typedef struct {
+    int cutoff;
+    char *pattern;
+    Patprog patprog;
+    char *guard;
+    char *action;
+    LinkList branches;
+} RParseState;
+
+typedef struct {
+  RParseState *state;
+  LinkList actions;
+} RParseBranch;
+
+typedef struct {
+    LinkList nullacts;
+    LinkList in;
+    LinkList out;
+} RParseResult;
+
+static char **rparseargs;
+static LinkList rparsestates;
+
+static int rparsealt(RParseResult *result, jmp_buf *perr);
+
+static void
+connectstates(LinkList out, LinkList in)
+{
+    LinkNode outnode, innode, ln;
+    for(outnode = firstnode(out); outnode; outnode = nextnode(outnode)) {
+	RParseBranch *outbranch = getdata(outnode);
+	for(innode = firstnode(in); innode; innode = nextnode(innode)) {
+	    RParseBranch *inbranch = getdata(innode);
+	    RParseBranch *br = ncalloc(sizeof(*br));
+	    br->state = inbranch->state;
+	    br->actions = newlinklist();
+	    for(ln = firstnode(outbranch->actions); ln; ln = nextnode(ln))
+	      addlinknode(br->actions, getdata(ln));
+	    for(ln = firstnode(inbranch->actions); ln; ln = nextnode(ln))
+	      addlinknode(br->actions, getdata(ln));
+	    addlinknode(outbranch->state->branches, br);
+	}
+    }
+}
+
+static int
+rparseelt(RParseResult *result, jmp_buf *perr)
+{
+    int l;
+    char *s = *rparseargs;
+
+    if(!s)
+        return 1;
+
+    switch(s[0]) {
+    case '/': {
+	RParseState *st;
+	RParseBranch *br;
+	char *pattern, *lookahead;
+	int patternlen, lookaheadlen;
+	l = strlen(s);
+	if(!((2 <= l && s[l - 1] == '/') ||
+	     (3 <= l && s[l - 2] == '/' && (s[l - 1] == '+' ||
+					    s[l - 1] == '-'))))
+	    return 1;
+	st = ncalloc(sizeof(*st));
+	st->branches = newlinklist();
+	st->cutoff = s[l - 1];
+	if(s[l - 1] == '/') {
+	    pattern = s + 1;
+	    patternlen = l - 2;
+	}
+	else {
+	    pattern = s + 1;
+	    patternlen = l - 3;
+	}
+	rparseargs++;
+	if((s = *rparseargs) && s[0] == '%' &&
+	   2 <= (l = strlen(s)) && s[l - 1] == '%') {
+	    rparseargs++;
+	    lookahead = s + 1;
+	    lookaheadlen = l - 2;
+	}
+	else {
+	    lookahead = NULL;
+	}
+	if(patternlen == 2 && !strncmp(pattern, "[]", 2))
+	    st->pattern = NULL;
+	else {
+	    char *cp;
+	    int l = patternlen + 12; /* (#b)((#B)...)...* */
+	    if(lookahead)
+	        l += lookaheadlen + 4; /* (#B)... */
+	    cp = st->pattern = ncalloc(l);
+	    strcpy(cp, "(#b)((#B)");
+	    cp += 9;
+	    strcpy(cp, pattern);
+	    cp += patternlen;
+	    strcpy(cp, ")");
+	    cp += 1;
+	    if(lookahead) {
+		strcpy(cp, "(#B)");
+		cp += 4;
+		strcpy(cp, lookahead);
+		cp += lookaheadlen;
+	    }
+	    strcpy(cp, "*");
+	}
+	st->patprog = NULL;
+	if((s = *rparseargs) && *s == '-') {
+	    rparseargs++;
+	    l = strlen(s);
+	    st->guard = ncalloc(l);
+	    memcpy(st->guard, s + 1, l - 1);
+	    st->guard[l - 1] = '\0';
+	}
+	else
+	    st->guard = NULL;
+	if((s = *rparseargs) && *s == ':') {
+	    rparseargs++;
+	    l = strlen(s);
+	    st->action = ncalloc(l);
+	    memcpy(st->action, s + 1, l - 1);
+	    st->action[l - 1] = '\0';
+	}
+	else
+	    st->action = NULL;
+	result->nullacts = NULL;
+	result->in = newlinklist();
+	br = ncalloc(sizeof(*br));
+	br->state = st;
+	br->actions = newlinklist();
+	addlinknode(result->in, br);
+	result->out = newlinklist();
+	br = ncalloc(sizeof(*br));
+	br->state = st;
+	br->actions = newlinklist();
+	addlinknode(result->out, br);
+	break;
+    }
+    case '(':
+	if(s[1])
+	    return 1;
+	rparseargs++;
+	if(rparsealt(result, perr))
+	    longjmp(*perr, 2);
+	s = *rparseargs;
+	if(!s || s[0] != ')' || s[1] != '\0')
+	    longjmp(*perr, 2);
+	rparseargs++;
+        break;
+    default:
+        return 1;
+    }
+
+    return 0;
+}
+
+static int
+rparseclo(RParseResult *result, jmp_buf *perr)
+{
+    if(rparseelt(result, perr))
+	return 1;
+
+    if(*rparseargs && !strcmp(*rparseargs, "#")) {
+	rparseargs++;
+	while(*rparseargs && !strcmp(*rparseargs, "#"))
+	    rparseargs++;
+
+	connectstates(result->out, result->in);
+	result->nullacts = newlinklist();
+    }
+    return 0;
+}
+
+static void
+prependactions(LinkList acts, LinkList branches)
+{
+    LinkNode aln, bln;
+    for(bln = firstnode(branches); bln; bln = nextnode(bln)) {
+	RParseBranch *br = getdata(bln);
+	for(aln = lastnode(acts); aln != (LinkNode)acts; aln = prevnode(aln))
+	    pushnode(br->actions, getdata(aln));
+    }
+}
+
+static void
+appendactions(LinkList acts, LinkList branches)
+{
+    LinkNode aln, bln;
+    for(bln = firstnode(branches); bln; bln = nextnode(bln)) {
+	RParseBranch *br = getdata(bln);
+	for(aln = firstnode(acts); aln; aln = nextnode(aln))
+	    addlinknode(br->actions, getdata(aln));
+    }
+}
+
+static int
+rparseseq(RParseResult *result, jmp_buf *perr)
+{
+    int l;
+    char *s;
+    RParseResult sub;
+
+    result->nullacts = newlinklist();
+    result->in = newlinklist();
+    result->out = newlinklist();
+
+    while(1) {
+	if((s = *rparseargs) && s[0] == '{' && s[(l = strlen(s)) - 1] == '}') {
+	    char *action = ncalloc(l - 1);
+	    LinkNode ln;
+	    rparseargs++;
+	    memcpy(action, s + 1, l - 2);
+	    action[l - 2] = '\0';
+	    if(result->nullacts)
+		addlinknode(result->nullacts, action);
+	    for(ln = firstnode(result->out); ln; ln = nextnode(ln)) {
+		RParseBranch *br = getdata(ln);
+		addlinknode(br->actions, action);
+	    }
+	}
+        else if(!rparseclo(&sub, perr)) {
+	    connectstates(result->out, sub.in);
+
+	    if(result->nullacts) {
+		prependactions(result->nullacts, sub.in);
+		insertlinklist(sub.in, lastnode(result->in), result->in);
+	    }
+	    if(sub.nullacts) {
+		appendactions(sub.nullacts, result->out);
+		insertlinklist(sub.out, lastnode(result->out), result->out);
+	    }
+	    else
+		result->out = sub.out;
+
+	    if(result->nullacts && sub.nullacts)
+		insertlinklist(sub.nullacts, lastnode(result->nullacts), result->nullacts);
+	    else
+		result->nullacts = NULL;
+	}
+	else
+	    break;
+    }
+    return 0;
+}
+
+static int
+rparsealt(RParseResult *result, jmp_buf *perr)
+{
+    RParseResult sub;
+
+    if(rparseseq(result, perr))
+	return 1;
+
+    while(*rparseargs && !strcmp(*rparseargs, "|")) {
+	rparseargs++;
+	if(rparseseq(&sub, perr))
+	    longjmp(*perr, 2);
+	if(!result->nullacts && sub.nullacts) {
+	    result->nullacts = sub.nullacts;
+	}
+	insertlinklist(sub.in, lastnode(result->in), result->in);
+	insertlinklist(sub.out, lastnode(result->out), result->out);
+    }
+    return 0;
+}
+
+static int
+rmatch(RParseResult *sm, char *subj, char *var1, char *var2)
+{
+    LinkNode ln, lnn;
+    LinkList nexts;
+    LinkList nextslist;
+    RParseBranch *br;
+    RParseState *st = NULL;
+    int point1 = 0, point2 = 0;
+
+    setiparam(var1, point1);
+    setiparam(var2, point2);
+
+    if(!*subj) {
+	if(sm->nullacts)
+	    for(ln = firstnode(sm->nullacts); ln; ln = nextnode(ln)) {
+	        char *action = getdata(ln);
+		if(action)
+		    execstring(action, 1, 0);
+	    }
+	return 0;
+    }
+
+    nextslist = newlinklist();
+    nexts = sm->in;
+    addlinknode(nextslist, nexts);
+    do {
+	char **savematch1, **savembegin1, **savemend1;
+	char **savematch2, **savembegin2, **savemend2;
+	PERMALLOC {
+	  savematch1 = duparray(getaparam("match"), (VFunc) dupstring);
+	  savembegin1 = duparray(getaparam("mbegin"), (VFunc) dupstring);
+	  savemend1 = duparray(getaparam("mend"), (VFunc) dupstring);
+	} LASTALLOC;
+	for(ln = firstnode(nexts); ln; ln = nextnode(ln)) {
+	    int i;
+	    RParseState *next;
+	    br = getdata(ln);
+	    next = br->state;
+	    if(next->pattern && !next->patprog) {
+	        tokenize(next->pattern);
+		if(!(next->patprog = patcompile(next->pattern, 0, NULL))) {
+		    return 2;
+		}
+	    }
+	    if(next->pattern && pattry(next->patprog, subj) &&
+	       (!next->guard || (execstring(next->guard, 1, 0), !lastval))) {
+		LinkNode aln;
+		char **mend = getaparam("mend");
+		int len = atoi(mend[0]);
+		for(i = len; i; i--)
+		  if(*subj++ == Meta)
+		    subj++;
+		PERMALLOC {
+		  savematch2 = duparray(getaparam("match"), (VFunc) dupstring);
+		  savembegin2 = duparray(getaparam("mbegin"), (VFunc) dupstring);
+		  savemend2 = duparray(getaparam("mend"), (VFunc) dupstring);
+		} LASTALLOC;
+		if(savematch1) setaparam("match", savematch1);
+		if(savembegin1) setaparam("mbegin", savembegin1);
+		if(savemend1) setaparam("mend", savemend1);
+		for(aln = firstnode(br->actions); aln; aln = nextnode(aln)) {
+		    char *action = getdata(aln);
+		    if(action)
+			execstring(action, 1, 0);
+		}
+		if(savematch2) setaparam("match", savematch2);
+		if(savembegin2) setaparam("mbegin", savembegin2);
+		if(savemend2) setaparam("mend", savemend2);
+		point2 += len;
+		setiparam(var2, point2);
+		st = br->state;
+		nexts = st->branches;
+		if(next->cutoff == '-' || (next->cutoff == '/' && len)) {
+		  nextslist = newlinklist();
+		  point1 = point2;
+		  setiparam(var1, point1);
+		}
+		addlinknode(nextslist, nexts);
+		break;
+	    }
+	}
+	if(!ln) {
+	    if(savematch1) freearray(savematch1);
+	    if(savembegin1) freearray(savembegin1);
+	    if(savemend1) freearray(savemend1);
+	}
+    } while(ln);
+
+    if(!*subj)
+        for(ln = firstnode(sm->out); ln; ln = nextnode(ln)) {
+	    br = getdata(ln);
+	    if(br->state == st) {
+		for(ln = firstnode(br->actions); ln; ln = nextnode(ln)) {
+		    char *action = getdata(ln);
+		    if(action)
+			execstring(action, 1, 0);
+		}
+		return 0;
+	    }
+	}
+
+    for(lnn = firstnode(nextslist); lnn; lnn = nextnode(lnn)) {
+	nexts = getdata(lnn);
+	for(ln = firstnode(nexts); ln; ln = nextnode(ln)) {
+	    br = getdata(ln);
+	    if(br->state->action)
+		execstring(br->state->action, 1, 0);
+	}
+    }
+    return empty(nexts) ? 2 : 1;
+}
+
+/*
+  usage: regexparse string regex...
+  status:
+    0: matched
+    1: unmatched (all next state candidates are failed)
+    2: unmatched (there is no next state candidates)
+    3: regex parse error
+*/
+
+static int
+bin_regexparse(char *nam, char **args, char *ops, int func)
+{
+    int oldextendedglob = opts[EXTENDEDGLOB];
+    char *var1 = args[0];
+    char *var2 = args[1];
+    char *subj = args[2];
+    int ret;
+    jmp_buf rparseerr;
+    RParseResult result;
+
+    opts[EXTENDEDGLOB] = 1;
+
+    rparseargs = args + 3;
+    HEAPALLOC {
+	pushheap();
+        rparsestates = newlinklist();
+	if(setjmp(rparseerr) || rparsealt(&result, &rparseerr) || *rparseargs) {
+	    if(*rparseargs)
+		zwarnnam(nam, "invalid regex : %s", *rparseargs, 0);
+	    else
+		zwarnnam(nam, "not enough regex arguments", NULL, 0);
+	    ret = 3;
+	}
+	else
+	    ret = 0;
+
+	if(!ret)
+	    ret = rmatch(&result, subj, var1, var2);
+        popheap();
+    } LASTALLOC;
+
+    opts[EXTENDEDGLOB] = oldextendedglob;
+    return ret;
+}
+
 static struct builtin bintab[] = {
     BUILTIN("zstyle", 0, bin_zstyle, 0, -1, 0, NULL, NULL),
     BUILTIN("zformat", 0, bin_zformat, 3, -1, 0, NULL, NULL),
+    BUILTIN("regexparse", 0, bin_regexparse, 3, -1, 0, NULL, NULL),
 };
 
 
-- 
Tanaka Akira