diff options
| -rw-r--r-- | ChangeLog | 5 | ||||
| -rw-r--r-- | string.c | 25 | ||||
| -rw-r--r-- | test/ruby/test_string.rb | 7 |
3 files changed, 31 insertions, 6 deletions
@@ -1,3 +1,8 @@ +Sun Feb 22 14:33:06 2009 Nobuyoshi Nakada <nobu@ruby-lang.org> + + * string.c (tr_trans): should not be affected by the encoding of + replacement unless actually modified. [ruby-talk:328967] + Sun Feb 22 13:38:44 2009 Nobuyoshi Nakada <nobu@ruby-lang.org> * ext/socket/init.c (s_recvfrom_nonblock): handles EAGAIN too. @@ -4791,8 +4791,10 @@ tr_trans(VALUE str, VALUE src, VALUE repl, int sflag) char *buf = ALLOC_N(char, max), *t = buf; while (s < send) { - c0 = c = rb_enc_codepoint(s, send, enc); - tlen = clen = rb_enc_codelen(c, enc); + int may_modify = 0; + c0 = c = rb_enc_codepoint(s, send, e1); + clen = rb_enc_codelen(c, e1); + tlen = enc == e1 ? clen : rb_enc_codelen(c, enc); s += clen; if (c < 256) { @@ -4819,6 +4821,7 @@ tr_trans(VALUE str, VALUE src, VALUE repl, int sflag) else { save = -1; c = c0; + if (enc != e1) may_modify = 1; } while (t - buf + tlen >= max) { offset = t - buf; @@ -4827,6 +4830,9 @@ tr_trans(VALUE str, VALUE src, VALUE repl, int sflag) t = buf + offset; } rb_enc_mbcput(c, t, enc); + if (may_modify && memcmp(s, t, tlen) != 0) { + modify = 1; + } t += tlen; } *t = '\0'; @@ -4858,8 +4864,10 @@ tr_trans(VALUE str, VALUE src, VALUE repl, int sflag) char *buf = ALLOC_N(char, max), *t = buf; while (s < send) { - c0 = c = rb_enc_codepoint(s, send, enc); - tlen = clen = rb_enc_codelen(c, enc); + int may_modify = 0; + c0 = c = rb_enc_codepoint(s, send, e1); + clen = rb_enc_codelen(c, e1); + tlen = enc == e1 ? clen : rb_enc_codelen(c, enc); if (c < 256) { c = trans[c]; @@ -4881,8 +4889,8 @@ tr_trans(VALUE str, VALUE src, VALUE repl, int sflag) modify = 1; } else { - modify = 1; c = c0; + if (enc != e1) may_modify = 1; } while (t - buf + tlen >= max) { offset = t - buf; @@ -4890,7 +4898,12 @@ tr_trans(VALUE str, VALUE src, VALUE repl, int sflag) REALLOC_N(buf, char, max); t = buf + offset; } - if (s != t) rb_enc_mbcput(c, t, enc); + if (s != t) { + rb_enc_mbcput(c, t, enc); + if (may_modify && memcmp(s, t, tlen) != 0) { + modify = 1; + } + } s += clen; t += tlen; } diff --git a/test/ruby/test_string.rb b/test/ruby/test_string.rb index eba0256439..4bee19bca3 100644 --- a/test/ruby/test_string.rb +++ b/test/ruby/test_string.rb @@ -1395,6 +1395,9 @@ class TestString < Test::Unit::TestCase assert_equal(S("hippo"), S("hello").tr(S("el"), S("ip"))) assert_equal(S("*e**o"), S("hello").tr(S("^aeiou"), S("*"))) assert_equal(S("hal"), S("ibm").tr(S("b-z"), S("a-z"))) + + a = "abc".force_encoding(Encoding::US_ASCII) + assert_equal(Encoding::US_ASCII, a.tr(S("z"), S("\u0101")).encoding) end def test_tr! @@ -1415,6 +1418,10 @@ class TestString < Test::Unit::TestCase a = S("ibm") assert_nil(a.tr!(S("B-Z"), S("A-Z"))) assert_equal(S("ibm"), a) + + a = "abc".force_encoding(Encoding::US_ASCII) + assert_nil(a.tr!(S("z"), S("\u0101"))) + assert_equal(Encoding::US_ASCII, a.encoding) end def test_tr_s |