--- /usr/local/lib/erlang/lib/stdlib-1.9.2/src/regexp.erl	Sat Feb 24 14:35:17 2001
+++ gregexp.erl	Thu Mar 22 12:43:51 2001
@@ -14,8 +14,18 @@
 %% AB. All Rights Reserved.''
 %% 
 %%     $Id$
+
+%% Submatch extraction (C) March 2001 pascal.brisset@cellicium.com .
+%% This module extends otp_src_R7B-1/libstdlib-1.9.2/src/regexp.erl
+%% with the syntax "\\(" and "\\)". This makes it possible to extract
+%% subgroups of a match. For example:
+%% URL="\\(.+\\)://\\(.+\\)\\(/.+\\)(\\?\\(.*\\)(&\\(.*\\))*)?",
+%%     gregexp:groups("http://localhost:81/script?arg&arg2&arg3", URL).
+%% {match,["http","localhost:81","/script","arg","arg2","arg3"]}
 %%
--module(regexp).
+%% Note that the character '(' can be matched with the regexp "[(]".
+
+-module(gregexp).
 
 %% This module provides a basic set of regular expression functions
 %% for strings. The functions provided are taken from AWK.
@@ -26,11 +36,13 @@
 
 -export([sh_to_awk/1,parse/1,format_error/1,match/2,first_match/2,matches/2]).
 -export([sub/3,gsub/3,split/2]).
+-export([groups/2]).
 
 -import(string, [substr/2,substr/3]).
 -import(lists, [reverse/1]).
 
 %% -type matchres() = {match,Start,Length} | nomatch | {error,E}.
+%% -type gmatchres() = {match,[string()]} | nomatch | {error,E}.
 %% -type subres() = {ok,RepString,RepCount} | {error,E}.
 %% -type splitres() = {ok,[SubString]} | {error,E}.
 
@@ -50,6 +62,7 @@
 %% reg3 -> reg3 "?" : {optional,'$1'}.
 %% reg3 -> reg4 : '$1'.
 %% reg4 -> "(" reg ")" : '$2'.
+%% reg4 -> "\\(" reg "\\)" : '$2'.
 %% reg4 -> "\\" char : '$2'.
 %% reg4 -> "^" : bos.
 %% reg4 -> "$" : eos.
@@ -85,6 +98,7 @@
     {L,S1} = reg3(S0),
     reg2p(S1, L).
 
+reg2p([$\\,$)|_]=S, L) -> {L,S};
 reg2p([C|S0], L) when C /= $|, C /= $) ->
     {R,S1} = reg3([C|S0]),
     reg2p(S1, {concat,L,R});
@@ -110,6 +124,11 @@
 	{R,[$)|S1]} -> {R,S1};
 	{R,S} -> throw({error,{unterminated,"("}})
     end;
+reg4([$\\,$(|S0]) ->
+    case reg(S0) of
+	{R,[$\\,$)|S1]} -> {{group,R},S1};
+	{R,S} -> throw({error,{unterminated,"\\("}})
+    end;
 reg4([$\\,O1,O2,O3|S]) when
   O1 >= $0, O1 =< $7, O2 >= $0, O2 =< $7, O3 >= $0, O3 =< $7 ->
     {(O1*8 + O2)*8 + O3 - 73*$0,S};
@@ -175,7 +194,7 @@
 %    char_string(S0, {concat,L,C});
 %char_string(S, L) -> {L,S}.
 
-%% -deftype re_app_res() = {match,RestPos,Rest} | nomatch.
+%% -deftype re_app_res() = {match,RestPos,Rest,Groups} | nomatch.
 
 %% re_apply(String, StartPos, RegExp) -> re_app_res().
 %%
@@ -207,6 +226,15 @@
 		re_apply(CE, [{kclosure,CE}|More], S, P));
 re_apply({pclosure,CE}, More, S, P) ->
     re_apply(CE, [{kclosure,CE}|More], S, P);
+re_apply({group,RE}, More, S, P) ->
+    %% Insert a pseudo-regexp so that we can record the group when we
+    %% reach its end.
+    re_apply(RE, [{endgroup,P}|More], S, P);
+re_apply({endgroup,St}, More, S, P) ->
+    case re_apply_more(More, S, P) of
+	nomatch -> nomatch;
+	{match, RP, R, G} -> {match, RP, R, [{St, P-St}|G]}
+    end;
 re_apply({optional,CE}, More, S, P) ->
     re_apply_or(re_apply_more(More, S, P),
 		re_apply(CE, More, S, P));
@@ -230,7 +258,7 @@
 %% re_apply_more([RegExp], String, Length) -> re_app_res().
 
 re_apply_more([RE|More], S, P) -> re_apply(RE, More, S, P);
-re_apply_more([], S, P) -> {match,P,S}.
+re_apply_more([], S, P) -> {match,P,S,[]}.
 
 %% in_char_class(Char, Class) -> bool().
 
@@ -243,8 +271,8 @@
 %%  If we want the best match then choose the longest match, else just
 %%  choose one by trying sequentially.
 
-re_apply_or({match,P1,S1}, {match,P2,S2}) when P1 >= P2 -> {match,P1,S1};
-re_apply_or({match,P1,S1}, {match,P2,S2}) -> {match,P2,S2};
+re_apply_or({match,P1,S1,G1},{match,P2,S2,G2}) when P1>=P2 -> {match,P1,S1,G1};
+re_apply_or({match,P1,S1,G1}, {match,P2,S2,G2}) -> {match,P2,S2,G2};
 re_apply_or(nomatch, R2) -> R2;
 re_apply_or(R1, nomatch) -> R1.
 
@@ -357,7 +385,7 @@
 
 first_match(RE, S, St) when S /= [] ->
     case re_apply(S, St, RE) of
-	{match,P,Rest} -> {St,P-St};
+	{match,P,Rest,_Groups} -> {St,P-St};
 	nomatch -> first_match(RE, tl(S), St+1)
     end;
 first_match(RE, [], St) -> nomatch.
@@ -446,10 +474,24 @@
 split_apply([], P, RE, T, Sub) -> [reverse(Sub)];
 split_apply(S, P, RE, T, Sub) ->
     case re_apply(S, P, RE) of
-	{match,P,Rest} ->
+	{match,P,Rest,_Groups} ->
 	    split_apply(tl(S), P+1, RE, T, [hd(S)|Sub]);
-	{match,P1,Rest} ->
+	{match,P1,Rest,_Groups} ->
 	    [reverse(Sub)|split_apply(Rest, P1, RE, T, [])];
 	nomatch ->
 	    split_apply(tl(S), P+1, RE, T, [hd(S)|Sub])
+    end.
+
+%%%%
+
+groups(S, RegExp) when list(RegExp) ->
+    {ok, ParsedRegExp} = parse(RegExp),
+    groups(S, ParsedRegExp);
+
+groups(S, ParsedRegExp) ->
+    case re_apply(S, 1, ParsedRegExp) of
+	{match, _RestPos, _Rest, Groups} ->
+	    GetGroup = fun ({Start,Len}) -> lists:sublist(S,Start,Len) end,
+	    {match, lists:map(GetGroup, Groups)};
+	Other -> Other
     end.