%%
%% %CopyrightBegin%
%%
%% Copyright Ericsson AB 1997-2009. All Rights Reserved.
%%
%% Licensed under the Apache License, Version 2.0 (the "License");
%% you may not use this file except in compliance with the License.
%% You may obtain a copy of the License at
%%
%% http://www.apache.org/licenses/LICENSE-2.0
%%
%% Unless required by applicable law or agreed to in writing, software
%% distributed under the License is distributed on an "AS IS" BASIS,
%% WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
%% See the License for the specific language governing permissions and
%% limitations under the License.
%%
%% %CopyrightEnd%
%%
%%
-module(icscan).
-export([scan/2]).
-include("ic.hrl").
%%----------------------------------------------------------------------
%%----------------------------------------------------------------------
-import(lists, [reverse/1]).
scan(G, File) ->
PL = call_preproc(G, File),
call_scan(G, PL).
call_preproc(G, File) ->
case ic_options:get_opt(G, use_preproc) of
true ->
icpreproc:preproc(G, File);
false ->
case catch file:read_file(File) of
{ok, Bin} ->
binary_to_list(Bin);
Other ->
exit(Other)
end
end.
call_scan(G, PL) ->
BE = ic_options:get_opt(G, be),
RSL = scan(G, BE, PL, 1, []),
lists:reverse(RSL).
%% Guard macros used at top scan functions only
-define(is_number(X), X >= $0 , X =< $9).
-define(is_upper(X), X >= $A , X =< $Z).
-define(is_lower(X), X >= $a, X =< $z).
-define(is_hex_uc(X), X >= $A , X =< $F).
-define(is_hex_lc(X), X >= $a , X =< $f).
-define(is_octal(X), X >=$0, X =< $7).
%% Handle:
%% const wchar aWChar = L'X';
scan(G, BE, [$L, $'|Str], Line, Out) ->
scan_const(G, BE, wchar, Str, [], Line, Out);
scan(G, BE, [$L, $"|Str], Line, Out) ->
scan_const(G, BE, wstring, Str, [], Line, Out);
scan(G, BE, [$_, X|Str], Line, Out) when ?is_upper(X) ->
scan_name(G, BE, Str, [X], false, Line, Out);
scan(G, BE, [$_, X|Str], Line, Out) when ?is_lower(X) ->
scan_name(G, BE, Str, [X], false, Line, Out);
scan(G, BE, [X|Str], Line, Out) when ?is_upper(X) ->
scan_name(G, BE, Str, [X], true, Line, Out);
scan(G, BE, [X|Str], Line, Out) when ?is_lower(X) ->
scan_name(G, BE, Str, [X], true, Line, Out);
scan(G, BE, [X|Str], Line, Out) when ?is_number(X) ->
scan_number(G, BE, Str, [X], Line, Out);
scan(G, BE, [9| T], Line, Out) -> scan(G, BE, T, Line, Out);
scan(G, BE, [32| T], Line, Out) -> scan(G, BE, T, Line, Out);
scan(G, BE, [$\r|Str], Line, Out) ->
scan(G, BE, Str, Line, Out);
scan(G, BE, [$\n|Str], Line, Out) ->
scan(G, BE, Str, Line+1, Out);
scan(G, BE, [$:, $: | Str], Line, Out) ->
scan(G, BE, Str, Line, [{'::', Line} | Out]);
scan(G, BE, [$/, $/ | Str], Line, Out) ->
Rest = skip_to_nl(Str),
scan(G, BE, Rest, Line, Out);
scan(G, BE, [$/, $* | Str], Line, Out) ->
Rest = skip_comment(Str),
scan(G, BE, Rest, Line, Out);
scan(G, BE, [$", $\\|Str], Line, Out) ->
scan_const(G, BE, string, [$\\|Str], [], Line, Out);
scan(G, BE, [$"|Str], Line, Out) ->
scan_const(G, BE, string, Str, [], Line, Out);
scan(G, BE, [$', $\\|Str], Line, Out) ->
scan_const(G, BE, char, [$\\|Str], [], Line, Out);
scan(G, BE, [$'|Str], Line, Out) ->
scan_const(G, BE, char, Str, [], Line, Out);
scan(G, BE, [$\\|Str], Line, Out) ->
scan_const(G, BE, escaped, [$\\|Str], [], Line, Out);
scan(G, BE, [$. | Str], Line, Out) ->
scan_frac(G, BE, Str, [$.], Line, Out);
scan(G, BE, [$# | Str], Line, Out) ->
scan_preproc(G, BE, Str, Line, Out);
scan(G, BE, [$<, $< | Str], Line, Out) ->
scan(G, BE, Str, Line, [{'<<', Line} | Out]);
scan(G, BE, [$>, $> | Str], Line, Out) ->
scan(G, BE, Str, Line, [{'>>', Line} | Out]);
scan(G, BE, [C|Str], Line, Out) ->
scan(G, BE, Str, Line, [{list_to_atom([C]), Line} | Out]);
scan(_G, _BE, [], _Line, Out) ->
Out.
scan_number(G, BE, [X|Str], [$0], Line, Out) when X == $X ; X ==$x ->
case Str of
[D|_TmpStr] when ?is_number(D); ?is_hex_uc(D); ?is_hex_lc(D) ->
{Num,Rest} = scan_hex_number(Str,0),
scan(G, BE, Rest, Line, [{'<integer_literal>', Line,
integer_to_list(Num)} | Out]);
[D|TmpStr] ->
scan(G, BE, TmpStr, Line, [{list_to_atom([D]), Line} | Out])
end;
scan_number(G, BE, Str, [$0], Line, Out) ->
%% If an integer literal starts with a 0 it may indicate that
%% it is represented as an octal number. But, it can also be a fixed
%% type which must use padding to match a fixed typedef. For example:
%% typedef fixed<5,2> fixed52;
%% 123.45d, 123.00d and 023.00d is all valid fixed values.
%% Naturally, a float can be defined as 0.14 or 00.14.
case pre_scan_number(Str, [], octal) of
octal ->
{Num, Rest} = scan_octal_number(Str,0),
scan(G, BE, Rest, Line, [{'<integer_literal>', Line,
integer_to_list(Num)} | Out]);
{fixed, Fixed, Rest} ->
scan(G, BE, Rest, Line, [{'<fixed_pt_literal>', Line, Fixed} | Out]);
float ->
%% Not very likely that someone defines a constant as 00.14 but ...
NewStr = remove_leading_zeroes(Str),
scan(G, BE, NewStr, Line, Out)
end;
scan_number(G, BE, [X|Str], Accum, Line, Out) when ?is_number(X) ->
scan_number(G, BE, Str, [X|Accum], Line, Out);
scan_number(G, BE, [X|Str], Accum, Line, Out) when X==$. ->
scan_frac(G, BE, Str, [X|Accum], Line, Out);
scan_number(G, BE, [X|Str], Accum, Line, Out) when X==$e ; X==$e ->
scan_exp(G, BE, Str, [X|Accum], Line, Out);
scan_number(G, BE, [X|Str], Accum, Line, Out) when X==$D ; X==$d ->
scan(G, BE, Str, Line, [{'<fixed_pt_literal>', Line,
(lists:reverse(Accum))} | Out]);
scan_number(G, BE, Str, Accum, Line, Out) ->
scan(G, BE, Str, Line, [{'<integer_literal>', Line,
(lists:reverse(Accum))} | Out]).
remove_leading_zeroes([$0|Rest]) ->
remove_leading_zeroes(Rest);
remove_leading_zeroes(L) ->
L.
scan_hex_number([X|Rest],Acc) when X >=$a, X =< $f ->
scan_hex_number(Rest,(Acc bsl 4) + (X - $a + 10));
scan_hex_number([X|Rest],Acc) when X >=$A, X =< $F ->
scan_hex_number(Rest,(Acc bsl 4) + (X - $A + 10));
scan_hex_number([X|Rest],Acc) when X >=$0, X =< $9 ->
scan_hex_number(Rest,(Acc bsl 4) + (X-$0));
scan_hex_number(Rest,Acc) ->
{Acc,Rest}.
pre_scan_number([$d|Rest], Acc, _) ->
{fixed, [$0|lists:reverse(Acc)], Rest};
pre_scan_number([$D|Rest], Acc, _) ->
{fixed, [$0|lists:reverse(Acc)], Rest};
pre_scan_number([$.|Rest], Acc, _) ->
%% Actually, we don't know if it's a float since it can be a fixed.
pre_scan_number(Rest, [$.|Acc], float);
pre_scan_number([X|_], _Acc, _) when X == $E ; X ==$e ->
%% Now we now it's a float.
float;
pre_scan_number([X|Rest], Acc, Type) when ?is_number(X) ->
pre_scan_number(Rest, [X|Acc], Type);
pre_scan_number(_Rest, _Acc, Type) ->
%% At this point we know it's a octal or float.
Type.
scan_octal_number([X|Rest],Acc) when ?is_octal(X) ->
scan_octal_number(Rest,(Acc bsl 3) + (X-$0));
scan_octal_number(Rest,Acc) ->
{Acc, Rest}.
%% Floating point number scan.
%%
%% Non trivial scan. A float consists of an integral part, a
%% decimal point, a fraction part, an e or E and a signed integer
%% exponent. Either the integer part or the fraction part but not
%% both may be missing, and either the decimal point or the
%% exponent part but not both may be missing. The exponent part
%% must consist of an e or E and a possibly signed exponent.
%%
%% Analysis shows that "1." ".7" "1e2" ".5e-3" "1.7e2" "1.7e-2"
%% is allowed and "1" ".e9" is not. The sign is only allowed just
%% after an e or E. The scanner reads a number as an integer
%% until it encounters a "." so the integer part only error case
%% will not be caught in the scanner (but rather in expression
%% evaluation)
scan_frac(G, _BE, [$e | _Str], [$.], Line, _Out) ->
ic_error:fatal_error(G, {illegal_float, Line});
scan_frac(G, _BE, [$E | _Str], [$.], Line, _Out) ->
ic_error:fatal_error(G, {illegal_float, Line});
scan_frac(G, BE, Str, Accum, Line, Out) ->
scan_frac2(G, BE, Str, Accum, Line, Out).
scan_frac2(G, BE, [X|Str], Accum, Line, Out) when ?is_number(X) ->
scan_frac2(G, BE, Str, [X|Accum], Line, Out);
scan_frac2(G, BE, [X|Str], Accum, Line, Out) when X==$e ; X==$E ->
scan_exp(G, BE, Str, [X|Accum], Line, Out);
%% The following case is for fixed (e.g. 123.45d).
scan_frac2(G, BE, [X|Str], Accum, Line, Out) when X==$d ; X==$D ->
scan(G, BE, Str, Line, [{'<fixed_pt_literal>', Line,
(lists:reverse(Accum))} | Out]);
scan_frac2(G, BE, Str, Accum, Line, Out) ->
scan(G, BE, Str, Line, [{'<floating_pt_literal>', Line,
(lists:reverse(Accum))} | Out]).
scan_exp(G, BE, [X|Str], Accum, Line, Out) when X==$- ->
scan_exp2(G, BE, Str, [X|Accum], Line, Out);
scan_exp(G, BE, Str, Accum, Line, Out) ->
scan_exp2(G, BE, Str, Accum, Line, Out).
scan_exp2(G, BE, [X|Str], Accum, Line, Out) when ?is_number(X) ->
scan_exp2(G, BE, Str, [X|Accum], Line, Out);
scan_exp2(G, BE, Str, Accum, Line, Out) ->
scan(G, BE, Str, Line, [{'<floating_pt_literal>', Line,
(lists:reverse(Accum))} | Out]).
scan_name(G, BE, [X|Str], Accum, TypeCheck, Line, Out) when ?is_upper(X) ->
scan_name(G, BE, Str, [X|Accum], TypeCheck, Line, Out);
scan_name(G, BE, [X|Str], Accum, TypeCheck, Line, Out) when ?is_lower(X) ->
scan_name(G, BE, Str, [X|Accum], TypeCheck, Line, Out);
scan_name(G, BE, [X|Str], Accum, TypeCheck, Line, Out) when ?is_number(X) ->
scan_name(G, BE, Str, [X|Accum], TypeCheck, Line, Out);
scan_name(G, BE, [$_|Str], Accum, TypeCheck, Line, Out) ->
scan_name(G, BE, Str, [$_|Accum], TypeCheck, Line, Out);
scan_name(G, BE, S, Accum, false, Line, Out) ->
%% The CORBA 2.3 specification allows the user to override typechecking:
%% typedef string _native;
%% interface i {
%% void foo(in _native VT);
%% };
%% BUT, the IFR-id remains the same ("IDL:native:1.0") etc. The reason for
%% this is that one don't have to re-write a large chunk of IDL- and
%% application-code.
scan(G, BE, S, Line, [{'<identifier>', Line, lists:reverse(Accum)} | Out]);
scan_name(G, BE, S, Accum, _, Line, Out) ->
L = lists:reverse(Accum),
X = case is_reserved(L, BE) of
undefined ->
{'<identifier>', Line, L};
Yes ->
{Yes, Line}
end,
scan(G, BE, S, Line, [X | Out]).
%% Shall scan a constant
scan_const(G, BE, string, [$" | Rest], Accum, Line, [{'<string_literal>', _, Str}|Out]) ->
scan(G, BE, Rest, Line,
[{'<string_literal>', Line, Str ++ lists:reverse(Accum)} | Out]);
scan_const(G, BE, string, [$" | Rest], Accum, Line, Out) ->
scan(G, BE, Rest, Line,
[{'<string_literal>', Line, lists:reverse(Accum)} | Out]);
scan_const(G, BE, wstring, [$" | Rest], Accum, Line, [{'<wstring_literal>', _,Wstr}|Out]) -> %% WSTRING
scan(G, BE, Rest, Line,
[{'<wstring_literal>', Line, Wstr ++ lists:reverse(Accum)} | Out]);
scan_const(G, BE, wstring, [$" | Rest], Accum, Line, Out) -> %% WSTRING
scan(G, BE, Rest, Line,
[{'<wstring_literal>', Line, lists:reverse(Accum)} | Out]);
scan_const(G, _BE, string, [], _Accum, Line, Out) -> %% Bad string
ic_error:error(G, {bad_string, Line}),
Out;
scan_const(G, _BE, wstring, [], _Accum, Line, Out) -> %% Bad WSTRING
ic_error:error(G, {bad_string, Line}),
Out;
scan_const(G, BE, char, [$' | Rest], Accum, Line, Out) ->
scan(G, BE, Rest, Line,
[{'<character_literal>', Line, lists:reverse(Accum)} | Out]);
scan_const(G, BE, wchar, [$' | Rest], Accum, Line, Out) -> %% WCHAR
scan(G, BE, Rest, Line,
[{'<wcharacter_literal>', Line, lists:reverse(Accum)} | Out]);
scan_const(G, BE, Mode, [$\\, C | Rest], Accum, Line, Out) ->
case escaped_char(C) of
error ->
ic_error:error(G, {bad_escape_character, Line, C}), %% Bad escape character
scan_const(G, BE, Mode, Rest, [C | Accum], Line, Out);
octal ->
{Num,Rest2} = scan_octal_number([C|Rest], 0),
scan_const(G, BE, Mode, Rest2, [Num|Accum], Line, Out);
hexadecimal ->
{Num,Rest2} = scan_hex_number(Rest, 0),
if
Num > 255 -> %% 16#FF
ic_error:error(G, {bad_escape_character, Line, C}),
scan_const(G, BE, Mode, Rest, [C | Accum], Line, Out);
true ->
scan_const(G, BE, Mode, Rest2, [Num|Accum], Line, Out)
end;
unicode ->
{Num,Rest2} = scan_hex_number(Rest, 0),
if
Num > 65535 -> %% 16#FFFF
ic_error:error(G, {bad_escape_character, Line, C}),
scan_const(G, BE, Mode, Rest, [C | Accum], Line, Out);
true ->
scan_const(G, BE, Mode, Rest2, [Num|Accum], Line, Out)
end;
EC ->
scan_const(G, BE, Mode, Rest, [EC | Accum], Line, Out)
end;
scan_const(G, BE, Mode, [C | Rest], Accum, Line, Out) ->
scan_const(G, BE, Mode, Rest, [C | Accum], Line, Out).
%%
%% Preprocessor output handling
%%
%% gcc outputs a line with line number, file name (within \") and
%% one or more integer flags. The scanner scans the line number,
%% the id and all integers up to nl.
%%
%% NOTE: This will have to be enhanced in order to eat #pragma
%%
scan_preproc(G, BE, Str, Line, Out) ->
{List, Rest} = scan_to_nl(strip(Str), []),
NewLine = get_new_line_nr(strip(List), Line+1, []),
case scan_number(G, BE, List, [], Line, [{'#', Line} | Out]) of
L when is_list(L) ->
scan(G, BE, Rest, NewLine, [{'#', Line} | L])
end.
get_new_line_nr([C|R], Line, Acc) when C>=$0, C=<$9 ->
get_new_line_nr(R, Line, [C|Acc]);
get_new_line_nr(_, Line, []) -> Line; % No line nr found
get_new_line_nr(_, _, Acc) -> list_to_integer(reverse(Acc)).
scan_to_nl([], Acc) -> {reverse(Acc), []};
scan_to_nl([$\n|Str], Acc) -> {reverse(Acc), Str};
scan_to_nl([$\r|R], Acc) -> scan_to_nl(R, Acc);
scan_to_nl([C|R], Acc) -> scan_to_nl(R, [C|Acc]).
strip([$ |R]) -> strip(R);
strip(L) -> L.
%% Escaped character. Escaped chars are repr as two characters in the
%% input list of letters and this is translated into one char.
escaped_char($n) -> $\n;
escaped_char($t) -> $\t;
escaped_char($v) -> $\v;
escaped_char($b) -> $\b;
escaped_char($r) -> $ ;
escaped_char($f) -> $\f;
escaped_char($a) -> $\a;
escaped_char($\\) -> $\\;
escaped_char($?) -> $?;
escaped_char($') -> $';
escaped_char($") -> $";
escaped_char($x) -> hexadecimal;
escaped_char($u) -> unicode;
escaped_char(X) when ?is_octal(X) -> octal;
%% Error
escaped_char(_Other) -> error.
skip_to_nl([]) -> [];
skip_to_nl([$\n | Str]) ->[$\n | Str];
skip_to_nl([_|Str]) ->
skip_to_nl(Str).
skip_comment([$\\, _ | Str]) ->
skip_comment(Str);
skip_comment([$*, $/ | Str]) -> Str;
skip_comment([_|Str]) ->
skip_comment(Str).
%%----------------------------------------------------------------------
%% Shall separate keywords from identifiers and numbers
%% Fill in the ets of reserved words
is_reserved("Object", _) -> 'Object';
is_reserved("in", _) -> in;
is_reserved("interface", _) -> interface;
is_reserved("case", _) -> 'case';
is_reserved("union", _) -> union;
is_reserved("struct", _) -> struct;
is_reserved("any", _) -> any;
is_reserved("long", _) -> long;
is_reserved("float", _) -> float;
is_reserved("out", _) -> out;
is_reserved("enum", _) -> enum;
is_reserved("double", _) -> double;
is_reserved("context", _) -> context;
is_reserved("oneway", _) -> oneway;
is_reserved("sequence", _) -> sequence;
is_reserved("FALSE", _) -> 'FALSE';
is_reserved("readonly", _) -> readonly;
is_reserved("char", _) -> char;
is_reserved("wchar", _) -> wchar;
is_reserved("void", _) -> void;
is_reserved("inout", _) -> inout;
is_reserved("attribute", _) -> attribute;
is_reserved("octet", _) -> octet;
is_reserved("TRUE", _) -> 'TRUE';
is_reserved("switch", _) -> switch;
is_reserved("unsigned", _) -> unsigned;
is_reserved("typedef", _) -> typedef;
is_reserved("const", _) -> const;
is_reserved("raises", _) -> raises;
is_reserved("string", _) -> string;
is_reserved("wstring", _) -> wstring;
is_reserved("default", _) -> default;
is_reserved("short", _) -> short;
is_reserved("module", _) -> module;
is_reserved("exception", _) -> exception;
is_reserved("boolean", _) -> boolean;
%% --- New keywords Introduced in CORBA-2.3.1 ---
%% For now we cannot add these for all backends right now since it would cause
%% some problems for at least one customer.
is_reserved("fixed", BE) -> check_be(BE, fixed);
%is_reserved("abstract", BE) -> check_be(BE, abstract);
%is_reserved("custom", BE) -> check_be(BE, custom);
%is_reserved("factory", BE) -> check_be(BE, factory);
%is_reserved("local", BE) -> check_be(BE, local);
%is_reserved("native", BE) -> check_be(BE, native);
%is_reserved("private", BE) -> check_be(BE, private);
%is_reserved("public", BE) -> check_be(BE, public);
%is_reserved("supports", BE) -> check_be(BE, supports);
%is_reserved("truncatable", BE) -> check_be(BE, truncatable);
%is_reserved("ValueBase", BE) -> check_be(BE, 'ValueBase');
%is_reserved("valuetype", BE) -> check_be(BE, valuetype);
is_reserved(_, _) -> undefined.
check_be(erl_corba, KeyWord) ->
KeyWord;
check_be(_, _) ->
undefined.
