diff options
| author | KOBAYASHI Shuji <[email protected]> | 2019-06-25 22:58:21 +0900 |
|---|---|---|
| committer | KOBAYASHI Shuji <[email protected]> | 2019-06-25 23:09:23 +0900 |
| commit | 75df13a97334c162b2cf743c3e37c4933a4b0d1c (patch) | |
| tree | 60968fa95933371218ab05d2493bd78d579e8445 /mrbgems/mruby-string-ext | |
| parent | cb3ee2d0501612f406e2d44b1e6d55b18861b1e1 (diff) | |
| download | mruby-75df13a97334c162b2cf743c3e37c4933a4b0d1c.tar.gz mruby-75df13a97334c162b2cf743c3e37c4933a4b0d1c.zip | |
Fix `String#byteslice` with `MRB_UTF8_STRING` and some edge cases
Example:
$ bin/mruby -e '
p "あa".byteslice(1)
p "bar".byteslice(3)
p "bar".byteslice(4..0)
'
Before this patch:
"a"
""
RangeError (4..0 out of range)
After this patch (same as Ruby):
"\x81"
nil
nil
Diffstat (limited to 'mrbgems/mruby-string-ext')
| -rw-r--r-- | mrbgems/mruby-string-ext/src/string.c | 58 | ||||
| -rw-r--r-- | mrbgems/mruby-string-ext/test/string.rb | 51 |
2 files changed, 74 insertions, 35 deletions
diff --git a/mrbgems/mruby-string-ext/src/string.c b/mrbgems/mruby-string-ext/src/string.c index d9ebb7392..50a4e5582 100644 --- a/mrbgems/mruby-string-ext/src/string.c +++ b/mrbgems/mruby-string-ext/src/string.c @@ -42,44 +42,32 @@ mrb_str_setbyte(mrb_state *mrb, mrb_value str) static mrb_value mrb_str_byteslice(mrb_state *mrb, mrb_value str) { - mrb_value a1; - mrb_int len; - - if (mrb_get_argc(mrb) == 2) { - mrb_int pos; - mrb_get_args(mrb, "ii", &pos, &len); - return mrb_str_substr(mrb, str, pos, len); + mrb_value a1, a2; + mrb_int str_len = RSTRING_LEN(str), beg, len; + mrb_bool empty = TRUE; + + if (mrb_get_args(mrb, "o|o", &a1, &a2) == 2) { + beg = mrb_fixnum(mrb_to_int(mrb, a1)); + len = mrb_fixnum(mrb_to_int(mrb, a2)); + goto subseq; } - mrb_get_args(mrb, "o|i", &a1, &len); - switch (mrb_type(a1)) { - case MRB_TT_RANGE: - { - mrb_int beg; - - len = RSTRING_LEN(str); - switch (mrb_range_beg_len(mrb, a1, &beg, &len, len, TRUE)) { - case MRB_RANGE_TYPE_MISMATCH: - break; - case MRB_RANGE_OK: - return mrb_str_substr(mrb, str, beg, len); - case MRB_RANGE_OUT: - mrb_raisef(mrb, E_RANGE_ERROR, "%S out of range", a1); - break; - } - return mrb_nil_value(); + if (mrb_type(a1) == MRB_TT_RANGE) { + if (mrb_range_beg_len(mrb, a1, &beg, &len, str_len, TRUE) == MRB_RANGE_OK) { + goto subseq; } -#ifndef MRB_WITHOUT_FLOAT - case MRB_TT_FLOAT: - a1 = mrb_fixnum_value((mrb_int)mrb_float(a1)); - /* fall through */ -#endif - case MRB_TT_FIXNUM: - return mrb_str_substr(mrb, str, mrb_fixnum(a1), 1); - default: - mrb_raise(mrb, E_TYPE_ERROR, "wrong type of argument"); + return mrb_nil_value(); + } + + beg = mrb_fixnum(mrb_to_int(mrb, a1)); + len = 1; + empty = FALSE; +subseq: + if (mrb_str_beg_len(str_len, &beg, &len) && (empty || len != 0)) { + return mrb_str_byte_subseq(mrb, str, beg, len); + } + else { + return mrb_nil_value(); } - /* not reached */ - return mrb_nil_value(); } /* diff --git a/mrbgems/mruby-string-ext/test/string.rb b/mrbgems/mruby-string-ext/test/string.rb index 02777e594..bf633bcef 100644 --- a/mrbgems/mruby-string-ext/test/string.rb +++ b/mrbgems/mruby-string-ext/test/string.rb @@ -26,10 +26,61 @@ end assert('String#byteslice') do str1 = "hello" + str2 = "\u3042ab" # "\xE3\x81\x82ab" + + assert_equal("h", str1.byteslice(0)) assert_equal("e", str1.byteslice(1)) + assert_equal(nil, str1.byteslice(5)) assert_equal("o", str1.byteslice(-1)) + assert_equal(nil, str1.byteslice(-6)) + assert_equal("\xE3", str2.byteslice(0)) + assert_equal("\x81", str2.byteslice(1)) + assert_equal(nil, str2.byteslice(5)) + assert_equal("b", str2.byteslice(-1)) + assert_equal(nil, str2.byteslice(-6)) + + assert_equal("", str1.byteslice(0, 0)) + assert_equal(str1, str1.byteslice(0, 6)) + assert_equal("el", str1.byteslice(1, 2)) + assert_equal("", str1.byteslice(5, 1)) + assert_equal("o", str1.byteslice(-1, 6)) + assert_equal(nil, str1.byteslice(-6, 1)) + assert_equal(nil, str1.byteslice(0, -1)) + assert_equal("", str2.byteslice(0, 0)) + assert_equal(str2, str2.byteslice(0, 6)) + assert_equal("\x81\x82", str2.byteslice(1, 2)) + assert_equal("", str2.byteslice(5, 1)) + assert_equal("b", str2.byteslice(-1, 6)) + assert_equal(nil, str2.byteslice(-6, 1)) + assert_equal(nil, str2.byteslice(0, -1)) + assert_equal("ell", str1.byteslice(1..3)) assert_equal("el", str1.byteslice(1...3)) + assert_equal("h", str1.byteslice(0..0)) + assert_equal("", str1.byteslice(5..0)) + assert_equal("o", str1.byteslice(4..5)) + assert_equal(nil, str1.byteslice(6..0)) + assert_equal("", str1.byteslice(-1..0)) + assert_equal("llo", str1.byteslice(-3..5)) + assert_equal("\x81\x82a", str2.byteslice(1..3)) + assert_equal("\x81\x82", str2.byteslice(1...3)) + assert_equal("\xE3", str2.byteslice(0..0)) + assert_equal("", str2.byteslice(5..0)) + assert_equal("b", str2.byteslice(4..5)) + assert_equal(nil, str2.byteslice(6..0)) + assert_equal("", str2.byteslice(-1..0)) + assert_equal("\x82ab", str2.byteslice(-3..5)) + + assert_raise(ArgumentError) { str1.byteslice } + assert_raise(ArgumentError) { str1.byteslice(1, 2, 3) } + assert_raise(TypeError) { str1.byteslice("1") } + assert_raise(TypeError) { str1.byteslice("1", 2) } + assert_raise(TypeError) { str1.byteslice(1, "2") } + assert_raise(TypeError) { str1.byteslice(1..2, 3) } + + skip unless Object.const_defined?(:Float) + assert_equal("o", str1.byteslice(4.0)) + assert_equal("\x82ab", str2.byteslice(2.0, 3.0)) end assert('String#dump') do |
