xmerl_regexp.erl

OTP是开放电信平台的简称

	{error,E} -> {error,E}
    end;
matches(S, {regexp,RE}) -> {match,matches_re(S, RE, 1)};
matches(S, {comp_regexp,RE}) -> {match,matches_comp(S, RE, 1)}.

matches_re([_|Cs]=S0, RE, P0) ->
    case re_apply(S0, P0, RE) of
	{match,P0,S1,_Subs} ->			%0 length match
	    [{P0,0}|matches_re(tl(S1), RE, P0+1)];
	{match,P1,S1,_Subs} ->
	    [{P0,P1-P0}|matches_re(S1, RE, P1)];
	nomatch -> matches_re(Cs, RE, P0+1);
	never_match -> []
    end;
matches_re([], _RE, _P) -> [].

matches_comp([_|Cs]=S0, RE, P0) ->
    case comp_apply(S0, P0, RE) of
	{match,P0,S1} ->			%0 length match
	    [{P0,0}|matches_comp(tl(S1), RE, P0+1)];
	{match,P1,S1} ->
	    [{P0,P1-P0}|matches_comp(S1, RE, P1)];
	nomatch -> matches_comp(Cs, RE, P0+1)
    end;
matches_comp([], _RE, _P) -> [].

%% sub(String, RegExp, Replace) -> {ok,RepString,RepCount} | {error,E}.
%%  Substitute the first match of the regular expression RegExp with
%%  the string Replace in String. Accept pre-parsed regular
%%  expressions.

sub(String, RegExp, Rep) when list(RegExp) ->
    case parse(RegExp) of
	{ok,RE} -> sub(String, RE, Rep);
	{error,E} -> {error,E}
    end;
sub(String, {regexp,RE}, Rep) ->
    case sub_re(String, 1, RE, [], Rep) of
	{yes,NewStr} -> {ok,NewStr,1};
	no -> {ok,String,0}
    end;
sub(String, {comp_regexp,RE}, Rep) ->
    case sub_comp(String, 1, RE, [], Rep) of
	{yes,NewStr} -> {ok,NewStr,1};
	no -> {ok,String,0}
    end.

%% sub_re(String, Position, Regexp, Before, Replacement) ->
%%      {NewString,Count}.
%% sub_comp(String, Position, Regexp, Before, Replacement) ->
%%      {NewString,Count}.
%% Step forward over String until a match is found saving stepped over
%% chars in Before. Return reversed Before prepended to replacement
%% and rest of string.

sub_re([C|Cs]=S0, P0, RE, Bef, Rep) ->
    case re_apply(S0, P0, RE) of
	{match,P0,_S1,_} ->			%Ignore 0 length match
	    sub_re(Cs, P0+1, RE, [C|Bef], Rep);
	{match,P1,Rest,_Gps} ->
	    {yes,reverse(Bef, sub_repl(Rep, substr(S0, 1, P1-P0), Rest))};
	nomatch -> sub_re(Cs, P0+1, RE, [C|Bef], Rep);
	never_match -> no			%No need to go on
    end;
sub_re([], _P, _RE, _Bef, _Rep) -> no.

sub_comp([C|Cs]=S0, P0, RE, Bef, Rep) ->
    case comp_apply(S0, P0, RE) of
	{match,P0,_S1} ->			%Ignore 0 length match
	    sub_comp(Cs, P0+1, RE, [C|Bef], Rep);
	{match,P1,Rest} ->
	    {yes,reverse(Bef, sub_repl(Rep, substr(S0, 1, P1-P0), Rest))};
	nomatch -> sub_comp(Cs, P0+1, RE, [C|Bef], Rep)
    end;
sub_comp([], _P, _RE, _Bef, _Rep) -> no.

sub_repl([$&|Rep], M, Rest) -> M ++ sub_repl(Rep, M, Rest);
sub_repl("\\&" ++ Rep, M, Rest) -> [$&|sub_repl(Rep, M, Rest)];
sub_repl([C|Rep], M, Rest) -> [C|sub_repl(Rep, M, Rest)];
sub_repl([], _M, Rest) -> Rest.

%%  gsub(String, RegExp, Replace) -> {ok,RepString,RepCount} | {error,E}.
%%  Substitute every match of the regular expression RegExp with the
%%  string New in String. Accept pre-parsed regular expressions.

gsub(String, RegExp, Rep) when list(RegExp) ->
    case parse(RegExp) of
	{ok,RE} -> gsub(String, RE, Rep);
	{error,E} -> {error,E}
    end;
gsub(String, {regexp,RE}, Rep) ->
    case gsub_re(String, 1, RE, [], Rep) of
	{NewStr,N} -> {ok,NewStr,N};
	no -> {ok,String,0}			%No substitutions
    end;
gsub(String, {comp_regexp,RE}, Rep) ->
    case gsub_comp(String, 1, RE, [], Rep) of
	{NewStr,N} -> {ok,NewStr,N};
	no -> {ok,String,0}			%No substitutions
    end.

%% gsub_re(String, Position, Regexp, Before, Replacement) ->
%%      {NewString,Count}.
%% gsub_comp(String, Position, Regexp, Before, Replacement) ->
%%      {NewString,Count}.
%% Step forward over String until a match is found saving stepped over
%% chars in Before. Call recursively to do rest of string after
%% match. Return reversed Before prepended to return from recursive
%% call.

gsub_re([C|Cs]=S0, P0, RE, Bef, Rep) ->
    case re_apply(S0, P0, RE) of
	{match,P0,_S1,_} ->			%Ignore 0 length match
	    gsub_re(Cs, P0+1, RE, [C|Bef], Rep);
	{match,P1,S1,_Gps} ->
	    case gsub_re(S1, P1, RE, [], Rep) of
		{NewStr,N0} ->			%Substituitions
		    {reverse(Bef, sub_repl(Rep, substr(S0, 1, P1-P0), NewStr)),
		     N0+1};
		no ->				%No substituitions.
		    {reverse(Bef, sub_repl(Rep, substr(S0, 1, P1-P0), S1)),1}
	    end;
	%%No match so step forward saving C on Bef.
	nomatch -> gsub_re(Cs, P0+1, RE, [C|Bef], Rep);
	never_match -> no			%No need to go on
    end;
gsub_re([], _P, _RE, _Bef, _Rep) -> no.

gsub_comp([C|Cs]=S0, P0, RE, Bef, Rep) ->
    case comp_apply(S0, P0, RE) of
	{match,P0,_S1} ->			%Ignore 0 length match
	    gsub_comp(Cs, P0+1, RE, [C|Bef], Rep);
	{match,P1,S1} ->
	    case gsub_comp(S1, P1, RE, [], Rep) of
		{NewStr,N0} ->			%Substituitions
		    {reverse(Bef, sub_repl(Rep, substr(S0, 1, P1-P0), NewStr)),
		     N0+1};
		no ->				%No substituitions.
		    {reverse(Bef, sub_repl(Rep, substr(S0, 1, P1-P0), S1)),1}
	    end;
	%%No match so step forward saving C on Bef.
	nomatch -> gsub_comp(Cs, P0+1, RE, [C|Bef], Rep)
    end;
gsub_comp([], _P, _RE, _Bef, _Rep) -> no.

%% split(String, RegExp) -> {ok,[SubString]} | {error,E}.
%%  Split a string into substrings where the RegExp describes the
%%  field seperator. The RegExp " " is specially treated.

split(String, " ") ->				%This is really special
    {ok,{regexp,RE}} = parse("[ \t]+"),
    case split_apply_re(String, RE, true) of
	[[]|Ss] -> {ok,Ss};
	Ss -> {ok,Ss}
    end;
split(String, RegExp) when list(RegExp) ->
    case parse(RegExp) of
	{ok,{regexp,RE}} -> {ok,split_apply_re(String, RE, false)};
	{error,E} -> {error,E}
    end;
split(String, {regexp,RE}) -> {ok,split_apply_re(String, RE, false)};
split(String, {comp_regexp,RE}) -> {ok,split_apply_comp(String, RE, false)}.

split_apply_re(S, RE, Trim) -> split_apply_re(S, 1, RE, Trim, []).

split_apply_re([], _P, _RE, true, []) -> [];
split_apply_re([], _P, _RE, _T, Sub) -> [reverse(Sub)];
split_apply_re([C|Cs]=S, P0, RE, T, Sub) ->
    case re_apply(S, P0, RE) of
	{match,P0,_S1,_} ->			%Ignore 0 length match
	    split_apply_re(Cs, P0+1, RE, T, [C|Sub]);
	{match,P1,S1,_} ->
	    [reverse(Sub)|split_apply_re(S1, P1, RE, T, [])];
	nomatch ->
	    split_apply_re(Cs, P0+1, RE, T, [C|Sub]);
	never_match -> [reverse(Sub, S)]	%No need to go on
    end.

split_apply_comp(S, RE, Trim) -> split_apply_comp(S, 1, RE, Trim, []).

%%split_apply_comp([], _P, _RE, true, []) -> [];
split_apply_comp([], _P, _RE, _T, Sub) -> [reverse(Sub)];
split_apply_comp([C|Cs]=S, P0, RE, T, Sub) ->
    case comp_apply(S, P0, RE) of
	{match,P0,_S1} ->			%Ignore 0 length match
	    split_apply_comp(Cs, P0+1, RE, T, [C|Sub]);
	{match,P1,S1} ->
	    [reverse(Sub)|split_apply_comp(S1, P1, RE, T, [])];
	nomatch ->
	    split_apply_comp(Cs, P0+1, RE, T, [C|Sub])
    end.

%% sub_match(String, RegExp) ->
%%      {match,Start,Length,SubExprs} | nomatch | {error,E}.
%%  Find the longest match of RegExp in String.

sub_match(S, RegExp) when list(RegExp) ->
    case parse(RegExp) of
	{ok,RE} -> sub_match(S, RE);
	{error,E} -> {error,E}
    end;
sub_match(S, {regexp,RE}) ->
    case sub_match_re(RE, S, 1, 0, -1, none) of
	{Start,Len,Subs} when Len >= 0 ->
	    {match,Start,Len,Subs};
	{_Start,_Len,_Subs} -> nomatch
    end.

sub_match_re(RE, S0, Pos0, Mst, Mlen, Msubs) ->
    case first_match_re(RE, S0, Pos0) of
	{St,Len,Subs} ->			%Found a match
	    Pos1 = St + 1,			%Where to start next match
	    S1 = lists:nthtail(Pos1-Pos0, S0),
	    if Len > Mlen -> sub_match_re(RE, S1, Pos1, St, Len, Subs);
	       true -> sub_match_re(RE, S1, Pos1, Mst, Mlen, Msubs)
	    end;
	nomatch -> {Mst,Mlen,Msubs}
    end.

%% sub_first_match(String, RegExp) ->
%%       {match,Start,Length,SubExprs} | nomatch | {error,E}.
%%  Find the longest match of RegExp in String, return Start and Length
%%  as well as tuple of sub-expression matches.

sub_first_match(S, RegExp) when is_list(RegExp) ->
    {ok,RE} = parse(RegExp),
    sub_first_match(S, RE);
sub_first_match(S, {regexp,RE}) ->
    case first_match_re(RE, S, 1) of
	{St,Len,Subs} -> {match,St,Len,Subs};
	nomatch -> nomatch
    end.


%% This is the regular expression grammar used. It is equivalent to the
%% one used in AWK, except that we allow ^ $ to be used anywhere and fail
%% in the matching.
%%
%% reg -> reg1 : '$1'.
%% reg1 -> reg1 "|" reg2 : {'or','$1','$2'}.
%% reg1 -> reg2 : '$1'.
%% reg2 -> reg2 reg3 : {concat,'$1','$2'}.
%% reg2 -> reg3 : '$1'.
%% reg3 -> reg3 "*" : {kclosure,'$1'}.
%% reg3 -> reg3 "+" : {pclosure,'$1'}.
%% reg3 -> reg3 "?" : {optional,'$1'}.
%% reg3 -> reg3 "{" [Min],[Max] "}" : {closure_range, Num, '$1'} see below
%% reg3 -> reg4 : '$1'.
%% reg4 -> "(" reg ")" : '$2'.
%% reg4 -> "\\" char : '$2'.
%% reg4 -> "^" : bos.
%% reg4 -> "$" : eos.
%% reg4 -> "." : char.
%% reg4 -> "[" class "]" : {char_class,char_class('$2')}
%% reg4 -> "[" "^" class "]" : {comp_class,char_class('$3')}
%% reg4 -> "\"" chars "\"" : char_string('$2')
%% reg4 -> char : '$1'.
%% reg4 -> empty : epsilon.
%%  The grammar of the current regular expressions. The actual parser
%%  is a recursive descent implementation of the grammar.

reg(S, Sc) -> reg1(S, Sc).

%% reg1 -> reg2 reg1'
%% reg1' -> "|" reg2
%% reg1' -> empty

reg1(S0, Sc0) ->
    {L,Sc1,S1} = reg2(S0, Sc0),
    reg1p(S1, L, Sc1).

reg1p([$||S0], L, Sc0) ->
    {R,Sc1,S1} = reg2(S0, Sc0),
    reg1p(S1, {'or',L,R}, Sc1);
reg1p(S, L, Sc) -> {L,Sc,S}.

%% reg2 -> reg3 reg2'
%% reg2' -> reg3
%% reg2' -> empty

reg2(S0, Sc0) ->
    {L,Sc1,S1} = reg3(S0, Sc0),
    reg2p(S1, L, Sc1).

reg2p([C|S0], L, Sc0) when C /= $|, C /= $) ->
    {R,Sc1,S1} = reg3([C|S0], Sc0),
    %% reg2p(S1, {concat,L,R}, Sc1);
    case is_integer(R) of
 	true -> 
 	    case L of
 		{literal,Lit} ->
 		    reg2p(S1, {literal,Lit ++[R]}, Sc1);
 		{concat,S2,Char} when is_integer(Char) ->
 		    reg2p(S1, {concat,S2,{literal,[Char,R]}}, Sc1);
 		{concat,S2,{literal,Lit}}  ->
 		    reg2p(S1, {concat,S2,{literal,Lit ++ [R]}}, Sc1);
 		Char when is_integer(Char) -> 
 		    reg2p(S1, {literal,[Char,R]}, Sc1);
 		_ ->
 		    reg2p(S1, {concat,L,R}, Sc1)
 	    end;
 	false ->
 	    reg2p(S1, {concat,L,R}, Sc1)
    end;
reg2p(S, L, Sc) -> {L,Sc,S}.

%% reg3 -> reg4 reg3'
%% reg3' -> "*" reg3'
%% reg3' -> "+" reg3'
%% reg3' -> "?" reg3'
%% reg3' -> "{" [Min],[Max] "}" reg3'
%% reg3' -> empty

reg3(S0, Sc0) ->
    {L,Sc1,S1} = reg4(S0, Sc0),
    reg3p(S1, L, Sc1).

reg3p([$*|S], L, Sc) -> reg3p(S, {kclosure,L}, Sc);
reg3p([$+|S], L, Sc) -> reg3p(S, {pclosure,L}, Sc);
reg3p([$?|S], L, Sc) -> reg3p(S, {optional,L}, Sc);
reg3p([${|Cs0], L, Sc) ->			% $}
    case interval_range(Cs0) of
	{none,none,_Cs1} -> parse_error({interval_range,[${|Cs0]});
	{N,M,[$}|Cs1]} -> reg3p(Cs1, {iclosure,L,N,M}, Sc);
	{_N,_M,_Cs1} -> parse_error({unterminated,"{"})
    end;
reg3p(S, L, Sc) -> {L,Sc,S}.

reg4([$(|S0], Sc0) ->
    Sc1 = Sc0+1,
    case reg(S0, Sc1) of
	{R,Sc2,[$)|S1]} -> {{subexpr,Sc1,R},Sc2,S1};
	{_R,_Sc,_S} -> parse_error({unterminated,"("})
    end;
reg4([$^|S], Sc) -> {bos,Sc,S};
reg4([$$|S], Sc) -> {eos,Sc,S};
reg4([$.|S], Sc) -> {{comp_class,"\n"},Sc,S};
reg4("[^" ++ S0, Sc) ->
    case char_class(S0) of
	{Cc,[$]|S1]} -> {{comp_class,Cc},Sc,S1};
	{_Cc,_S} -> parse_error({unterminated,"["})
    end;
reg4([$[|S0], Sc) ->
    case char_class(S0) of
	{Cc,[$]|S1]} -> {{char_class,Cc},Sc,S1};
	{_Cc,_S1} -> parse_error({unterminated,"["})
    end;
%reg4([$"|S0], Sc) ->
%    case char_string(S0) of
%	{St,[$"|S1]} -> {St,Sc,S1};
%	{St,S1} -> parse_error({unterminated,"\""})
%    end;
reg4([C0|S0], Sc) when
  is_integer(C0), C0 /= $*, C0 /= $+, C0 /= $?, C0 /= $], C0 /= $), C0 /= $} ->
    %% Handle \ quoted characters as well, at least those we see.
    {C1,S1} = char(C0, S0),
    {C1,Sc,S1};
reg4(S=[$)|_], Sc) -> {epsilon,Sc,S};
reg4([C|_S], _Sc) -> parse_error({illegal,[C]});