diff --git a/ChangeLog b/ChangeLog index 396bed0806..96fc45edfb 100644 --- a/ChangeLog +++ b/ChangeLog @@ -1,3 +1,8 @@ +Thu Dec 17 14:16:29 2015 Nobuyoshi Nakada + + * string.c (rb_str_scrub): the result should be infected by the + original string. + Thu Dec 17 13:35:27 2015 Nobuyoshi Nakada * transcode.c (rb_econv_substr_append, econv_primitive_convert): diff --git a/string.c b/string.c index 2c8089b94c..4ab89a8e2f 100644 --- a/string.c +++ b/string.c @@ -8667,6 +8667,10 @@ rb_str_scrub(VALUE str, VALUE repl) int cr = ENC_CODERANGE(str); rb_encoding *enc; int encidx; + VALUE buf = Qnil; + const char *rep; + long replen; + int tainted = 0; if (ENC_CODERANGE_CLEAN_P(cr)) return Qnil; @@ -8674,6 +8678,7 @@ rb_str_scrub(VALUE str, VALUE repl) enc = STR_ENC_GET(str); if (!NIL_P(repl)) { repl = str_compat_and_valid(repl, enc); + tainted = OBJ_TAINTED_RAW(repl); } if (rb_enc_dummy_p(enc)) { @@ -8690,10 +8695,7 @@ rb_str_scrub(VALUE str, VALUE repl) const char *p = RSTRING_PTR(str); const char *e = RSTRING_END(str); const char *p1 = p; - const char *rep; - long replen; int rep7bit_p; - VALUE buf = Qnil; if (rb_block_given_p()) { rep = NULL; replen = 0; @@ -8759,6 +8761,7 @@ rb_str_scrub(VALUE str, VALUE repl) else { repl = rb_yield(rb_enc_str_new(p, clen, enc)); repl = str_compat_and_valid(repl, enc); + tainted |= OBJ_TAINTED_RAW(repl); rb_str_buf_cat(buf, RSTRING_PTR(repl), RSTRING_LEN(repl)); if (ENC_CODERANGE(repl) == ENC_CODERANGE_VALID) cr = ENC_CODERANGE_VALID; @@ -8793,22 +8796,18 @@ rb_str_scrub(VALUE str, VALUE repl) else { repl = rb_yield(rb_enc_str_new(p, e-p, enc)); repl = str_compat_and_valid(repl, enc); + tainted |= OBJ_TAINTED_RAW(repl); rb_str_buf_cat(buf, RSTRING_PTR(repl), RSTRING_LEN(repl)); if (ENC_CODERANGE(repl) == ENC_CODERANGE_VALID) cr = ENC_CODERANGE_VALID; } } - ENCODING_CODERANGE_SET(buf, rb_enc_to_index(enc), cr); - return buf; } else { /* ASCII incompatible */ const char *p = RSTRING_PTR(str); const char *e = RSTRING_END(str); const char *p1 = p; - VALUE buf = Qnil; - const char *rep; - long replen; long mbminlen = rb_enc_mbminlen(enc); if (!NIL_P(repl)) { rep = RSTRING_PTR(repl); @@ -8863,6 +8862,7 @@ rb_str_scrub(VALUE str, VALUE repl) else { repl = rb_yield(rb_enc_str_new(p, e-p, enc)); repl = str_compat_and_valid(repl, enc); + tainted |= OBJ_TAINTED_RAW(repl); rb_str_buf_cat(buf, RSTRING_PTR(repl), RSTRING_LEN(repl)); } p += clen; @@ -8889,12 +8889,15 @@ rb_str_scrub(VALUE str, VALUE repl) else { repl = rb_yield(rb_enc_str_new(p, e-p, enc)); repl = str_compat_and_valid(repl, enc); + tainted |= OBJ_TAINTED_RAW(repl); rb_str_buf_cat(buf, RSTRING_PTR(repl), RSTRING_LEN(repl)); } } - ENCODING_CODERANGE_SET(buf, rb_enc_to_index(enc), ENC_CODERANGE_VALID); - return buf; + cr = ENC_CODERANGE_VALID; } + FL_SET_RAW(buf, tainted|OBJ_TAINTED_RAW(str)); + ENCODING_CODERANGE_SET(buf, rb_enc_to_index(enc), cr); + return buf; } /* diff --git a/test/ruby/test_m17n.rb b/test/ruby/test_m17n.rb index 74b847a825..9f96652083 100644 --- a/test/ruby/test_m17n.rb +++ b/test/ruby/test_m17n.rb @@ -1560,27 +1560,32 @@ class TestM17N < Test::Unit::TestCase assert_equal(a("\xE3\x81\x82"), s.b) assert_equal(Encoding::ASCII_8BIT, s.b.encoding) s.taint - assert_equal(true, s.b.tainted?) + assert_predicate(s.b, :tainted?) s = "abc".b - assert_equal(true, s.b.ascii_only?) + assert_predicate(s.b, :ascii_only?) end def test_scrub_valid_string str = "foo" assert_equal(str, str.scrub) assert_not_same(str, str.scrub) + assert_predicate(str.dup.taint.scrub, :tainted?) str = "\u3042\u3044" assert_equal(str, str.scrub) assert_not_same(str, str.scrub) + assert_predicate(str.dup.taint.scrub, :tainted?) str.force_encoding(Encoding::ISO_2022_JP) # dummy encoding assert_equal(str, str.scrub) assert_not_same(str, str.scrub) assert_nothing_raised(ArgumentError) {str.scrub(nil)} + assert_predicate(str.dup.taint.scrub, :tainted?) end def test_scrub_replace_default assert_equal("\uFFFD\uFFFD\uFFFD", u("\x80\x80\x80").scrub) assert_equal("\uFFFDA", u("\xF4\x80\x80A").scrub) + assert_predicate(u("\x80\x80\x80").taint.scrub, :tainted?) + assert_predicate(u("\xF4\x80\x80A").taint.scrub, :tainted?) # examples in Unicode 6.1.0 D93b assert_equal("\x41\uFFFD\uFFFD\x41\uFFFD\x41", @@ -1595,7 +1600,14 @@ class TestM17N < Test::Unit::TestCase def test_scrub_replace_argument assert_equal("foo", u("foo").scrub("\u3013")) + assert_predicate(u("foo").taint.scrub("\u3013"), :tainted?) + assert_not_predicate(u("foo").scrub("\u3013".taint), :tainted?) + assert_equal("\u3042\u3044", u("\xE3\x81\x82\xE3\x81\x84").scrub("\u3013")) + assert_predicate(u("\xE3\x81\x82\xE3\x81\x84").taint.scrub("\u3013"), :tainted?) + assert_not_predicate(u("\xE3\x81\x82\xE3\x81\x84").scrub("\u3013".taint), :tainted?) assert_equal("\u3042\u3013", u("\xE3\x81\x82\xE3\x81").scrub("\u3013")) + assert_predicate(u("\xE3\x81\x82\xE3\x81").taint.scrub("\u3013"), :tainted?) + assert_predicate(u("\xE3\x81\x82\xE3\x81").scrub("\u3013".taint), :tainted?) assert_raise(Encoding::CompatibilityError){ u("\xE3\x81\x82\xE3\x81").scrub(e("\xA4\xA2")) } assert_raise(TypeError){ u("\xE3\x81\x82\xE3\x81").scrub(1) } assert_raise(ArgumentError){ u("\xE3\x81\x82\xE3\x81\x82\xE3\x81").scrub(u("\x81")) } @@ -1604,6 +1616,8 @@ class TestM17N < Test::Unit::TestCase def test_scrub_replace_block assert_equal("\u3042", u("\xE3\x81\x82\xE3\x81").scrub{|x|'<'+x.unpack('H*')[0]+'>'}) + assert_predicate(u("\xE3\x81\x82\xE3\x81").taint.scrub{|x|'<'+x.unpack('H*')[0]+'>'}, :tainted?) + assert_predicate(u("\xE3\x81\x82\xE3\x81").scrub{|x|('<'+x.unpack('H*')[0]+'>').taint}, :tainted?) assert_raise(Encoding::CompatibilityError){ u("\xE3\x81\x82\xE3\x81").scrub{e("\xA4\xA2")} } assert_raise(TypeError){ u("\xE3\x81\x82\xE3\x81").scrub{1} } assert_raise(ArgumentError){ u("\xE3\x81\x82\xE3\x81\x82\xE3\x81").scrub{u("\x81")} }