aboutsummaryrefslogtreecommitdiffstats
path: root/lib/xmerl/src/xmerl_sax_parser_utf16be.erlsrc
diff options
context:
space:
mode:
Diffstat (limited to 'lib/xmerl/src/xmerl_sax_parser_utf16be.erlsrc')
-rw-r--r--lib/xmerl/src/xmerl_sax_parser_utf16be.erlsrc50
1 files changed, 46 insertions, 4 deletions
diff --git a/lib/xmerl/src/xmerl_sax_parser_utf16be.erlsrc b/lib/xmerl/src/xmerl_sax_parser_utf16be.erlsrc
index ff84ece97a..ec89024729 100644
--- a/lib/xmerl/src/xmerl_sax_parser_utf16be.erlsrc
+++ b/lib/xmerl/src/xmerl_sax_parser_utf16be.erlsrc
@@ -2,7 +2,7 @@
%%--------------------------------------------------------------------
%% %CopyrightBegin%
%%
-%% Copyright Ericsson AB 2008-2016. All Rights Reserved.
+%% Copyright Ericsson AB 2008-2017. 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.
@@ -34,8 +34,50 @@
-define(APPEND_STRING(Rest, New), <<Rest/binary, New/binary>>).
-define(TO_INPUT_FORMAT(Val), unicode:characters_to_binary(Val, unicode, {utf16, big})).
-%% STRING_REST and STRING_UNBOUND_REST is only different in the list case
-define(STRING_UNBOUND_REST(MatchChar, Rest), <<MatchChar/big-utf16, Rest/binary>>).
--define(BYTE_ORDER_MARK_1, undefined_bom1).
--define(BYTE_ORDER_MARK_2, <<16#FE>>).
+-define(BYTE_ORDER_MARK_1, <<16#FE>>).
-define(BYTE_ORDER_MARK_REST(Rest), <<16#FE, 16#FF, Rest/binary>>).
+
+-define(PARSE_BYTE_ORDER_MARK(Bytes, State),
+ parse_byte_order_mark(?STRING_EMPTY, State) ->
+ cf(?STRING_EMPTY, State, fun parse_byte_order_mark/2);
+ parse_byte_order_mark(?BYTE_ORDER_MARK_1, State) ->
+ cf(?BYTE_ORDER_MARK_1, State, fun parse_byte_order_mark/2);
+ parse_byte_order_mark(?BYTE_ORDER_MARK_REST(Rest), State) ->
+ parse_xml_decl(Rest, State);
+ parse_byte_order_mark(Bytes, State) ->
+ parse_xml_decl(Bytes, State)).
+
+-define(PARSE_XML_DECL(Bytes, State),
+ parse_xml_decl(Bytes, #xmerl_sax_parser_state{encoding=Enc} = State) when is_binary(Bytes) ->
+ case unicode:characters_to_list(Bytes, Enc) of
+ {incomplete, _, _} ->
+ cf(Bytes, State, fun parse_xml_decl/2);
+ {error, _Encoded, _Rest} ->
+ ?fatal_error(State, lists:flatten(io_lib:format("Bad character, not in ~p\n", [Enc])));
+ _ ->
+ parse_prolog(Bytes, State)
+ end;
+ parse_xml_decl(Bytes, State) ->
+ parse_prolog(Bytes, State)).
+
+-define(WHITESPACE(Bytes, State, Acc),
+ whitespace(?STRING_UNBOUND_REST(_C, _) = Bytes, State, Acc) ->
+ {lists:reverse(Acc), Bytes, State};
+ whitespace(Bytes, #xmerl_sax_parser_state{encoding=Enc} = State, Acc) when is_binary(Bytes) ->
+ case unicode:characters_to_list(Bytes, Enc) of
+ {incomplete, _, _} ->
+ cf(Bytes, State, Acc, fun whitespace/3);
+ {error, _Encoded, _Rest} ->
+ ?fatal_error(State, lists:flatten(io_lib:format("Bad character, not in ~p\n", [Enc])))
+ end).
+
+-define(PARSE_EXTERNAL_ENTITY_BYTE_ORDER_MARK(Bytes, State),
+ parse_external_entity_byte_order_mark(?STRING_EMPTY, State) ->
+ cf(?STRING_EMPTY, State, fun parse_external_entity_byte_order_mark/2);
+ parse_external_entity_byte_order_mark(?BYTE_ORDER_MARK_1, State) ->
+ cf(?BYTE_ORDER_MARK_1, State, fun parse_external_entity_byte_order_mark/2);
+ parse_external_entity_byte_order_mark(?BYTE_ORDER_MARK_REST(Rest), State) ->
+ parse_external_entity_1(Rest, State);
+ parse_external_entity_byte_order_mark(Bytes, State) ->
+ parse_external_entity_1(Bytes, State)).