Fix bug that allows invalid UTF-8 after valid escapes.
Merges r991073 from trunk to branches/1.0.x
Fixes COUCHDB-875
git-svn-id: https://svn.apache.org/repos/asf/couchdb/branches/1.0.x@1061088 13f79535-47bb-0310-9956-ffa450edef68
diff --git a/src/mochiweb/mochijson2.erl b/src/mochiweb/mochijson2.erl
index 111c37b..1ed2b88 100644
--- a/src/mochiweb/mochijson2.erl
+++ b/src/mochiweb/mochijson2.erl
@@ -405,8 +405,22 @@
Acc1 = lists:reverse(xmerl_ucs:to_utf8(C), Acc),
tokenize_string(B, ?ADV_COL(S, 6), Acc1)
end;
- <<_:O/binary, C, _/binary>> ->
- tokenize_string(B, ?INC_CHAR(S, C), [C | Acc])
+ <<_:O/binary, C1, _/binary>> when C1 < 128 ->
+ tokenize_string(B, ?INC_CHAR(S, C1), [C1 | Acc]);
+ <<_:O/binary, C1, C2, _/binary>> when C1 >= 194, C1 =< 223,
+ C2 >= 128, C2 =< 191 ->
+ tokenize_string(B, ?ADV_COL(S, 2), [C2, C1 | Acc]);
+ <<_:O/binary, C1, C2, C3, _/binary>> when C1 >= 224, C1 =< 239,
+ C2 >= 128, C2 =< 191,
+ C3 >= 128, C3 =< 191 ->
+ tokenize_string(B, ?ADV_COL(S, 3), [C3, C2, C1 | Acc]);
+ <<_:O/binary, C1, C2, C3, C4, _/binary>> when C1 >= 240, C1 =< 244,
+ C2 >= 128, C2 =< 191,
+ C3 >= 128, C3 =< 191,
+ C4 >= 128, C4 =< 191 ->
+ tokenize_string(B, ?ADV_COL(S, 4), [C4, C3, C2, C1 | Acc]);
+ _ ->
+ throw(invalid_utf8)
end.
tokenize_number(B, S) ->
@@ -653,7 +667,9 @@
<<?Q, 16#E0, 16#80,16#7F, ?Q>>,
<<?Q, 16#F0, 16#80, 16#80, 16#7F, ?Q>>,
% we don't support code points > 10FFFF per RFC 3629
- <<?Q, 16#F5, 16#80, 16#80, 16#80, ?Q>>
+ <<?Q, 16#F5, 16#80, 16#80, 16#80, ?Q>>,
+ %% escape characters trigger a different code path
+ <<?Q, $\\, $\n, 16#80, ?Q>>
],
lists:foreach(fun(X) ->
ok = try decode(X) catch invalid_utf8 -> ok end
