diff options
author | nobu <nobu@b2dd03c8-39d4-4d8f-98ff-823fe69b080e> | 2015-09-16 05:50:00 +0000 |
---|---|---|
committer | nobu <nobu@b2dd03c8-39d4-4d8f-98ff-823fe69b080e> | 2015-09-16 05:50:00 +0000 |
commit | 965d8241c1ac3556685a833206d644bd6c7dc646 (patch) | |
tree | 0cd1be11831d65517b08c912815a33d230f2948e | |
parent | 22190b3f32010075c1dacd6d9cbec3fa488a6187 (diff) | |
download | ruby-965d8241c1ac3556685a833206d644bd6c7dc646.tar.gz |
string.c: keep coderange
* string.c (rb_str_setbyte): keep the code range as possible.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@51873 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
-rw-r--r-- | ChangeLog | 4 | ||||
-rw-r--r-- | string.c | 45 | ||||
-rw-r--r-- | test/ruby/test_m17n.rb | 25 |
3 files changed, 72 insertions, 2 deletions
@@ -1,3 +1,7 @@ +Wed Sep 16 14:49:58 2015 Nobuyoshi Nakada <nobu@ruby-lang.org> + + * string.c (rb_str_setbyte): keep the code range as possible. + Wed Sep 16 13:23:48 2015 NAKAMURA Usaku <usa@ruby-lang.org> * doc/syntax/literals.rdoc (Strings): mention about ?a literal. @@ -4702,15 +4702,56 @@ rb_str_setbyte(VALUE str, VALUE index, VALUE value) long pos = NUM2LONG(index); int byte = NUM2INT(value); long len = RSTRING_LEN(str); + char *head, *ptr, *left = 0; + rb_encoding *enc; + int cr = ENC_CODERANGE_UNKNOWN, width, nlen; if (pos < -len || len <= pos) rb_raise(rb_eIndexError, "index %ld out of string", pos); if (pos < 0) pos += len; - rb_str_modify(str); + if (!str_independent(str)) + str_make_independent(str); + enc = STR_ENC_GET(str); + head = RSTRING_PTR(str); + ptr = &head[pos]; + if (len > RSTRING_EMBED_LEN_MAX) { + cr = ENC_CODERANGE(str); + switch (cr) { + case ENC_CODERANGE_7BIT: + left = ptr; + width = 1; + break; + case ENC_CODERANGE_VALID: + left = rb_enc_left_char_head(head, ptr, head+len, enc); + width = rb_enc_precise_mbclen(left, head+len, enc); + break; + default: + ENC_CODERANGE_CLEAR(str); + } + } + else { + ENC_CODERANGE_CLEAR(str); + } + + *ptr = byte; - RSTRING_PTR(str)[pos] = byte; + switch (cr) { + case ENC_CODERANGE_7BIT: + if (ISASCII(byte)) break; + case ENC_CODERANGE_VALID: + nlen = rb_enc_precise_mbclen(left, head+len, enc); + if (!MBCLEN_CHARFOUND_P(nlen)) + ENC_CODERANGE_SET(str, ENC_CODERANGE_BROKEN); + else if (cr == ENC_CODERANGE_7BIT) + ENC_CODERANGE_SET(str, ENC_CODERANGE_VALID); + else if (MBCLEN_CHARFOUND_LEN(nlen) != width) + ENC_CODERANGE_CLEAR(str); + else if (ISASCII(byte)) /* may become 7BIT */ + ENC_CODERANGE_CLEAR(str); + break; + } return value; } diff --git a/test/ruby/test_m17n.rb b/test/ruby/test_m17n.rb index 6b89ad572d..ca25f8502d 100644 --- a/test/ruby/test_m17n.rb +++ b/test/ruby/test_m17n.rb @@ -1488,6 +1488,31 @@ class TestM17N < Test::Unit::TestCase s = u("\xE3\x81\x82\xE3\x81\x84") s.setbyte(-4, 0x84) assert_equal(u("\xE3\x81\x84\xE3\x81\x84"), s) + + x = "x" * 100 + t = nil + failure = proc {"#{i}: #{encdump(t)}"} + + s = "\u{3042 3044}" + s.bytesize.times {|i| + t = s + x + t.setbyte(i, t.getbyte(i)+1) + assert_predicate(t, :valid_encoding?, failure) + assert_not_predicate(t, :ascii_only?, failure) + t = s + x + t.setbyte(i, 0x20) + assert_not_predicate(t, :valid_encoding?, failure) + } + + s = "\u{41 42 43}" + s.bytesize.times {|i| + t = s + x + t.setbyte(i, 0x20) + assert_predicate(t, :valid_encoding?, failure) + assert_predicate(t, :ascii_only?, failure) + t.setbyte(i, 0xe3) + assert_not_predicate(t, :valid_encoding?, failure) + } end def test_compatible |