Project

General

Profile

Download (8.95 KB)

Feature #12275 » v1.patch

tad (Tadashi Saito), 11/27/2017 07:56 PM

View differences:

NEWS
* String#each_grapheme_cluster and String#grapheme_clusters is added to
enumerate grapheme clusters [Feature #13780]
* String#start_with? supports regexp [Feature #13712]
* String#undump is added to unescape String#dump'ed string [Feature #12275]
* Regexp/String: Update Unicode version from 9.0.0 to 10.0.0 [Feature #13685]
string.c
#include "ruby_assert.h"
#include "id.h"
#include "debug_counter.h"
#include "ruby/util.h"
#define BEG(no) (regs->beg[(no)])
#define END(no) (regs->end[(no)])
......
return rb_str_eql(folded_str1, folded_str2);
}
static long
strseq_core(const char *str_ptr, const char *str_ptr_end, long str_len,
const char *sub_ptr, long sub_len, long offset, rb_encoding *enc)
{
const char *search_start = str_ptr;
long pos, search_len = str_len - offset;
for (;;) {
const char *t;
pos = rb_memsearch(sub_ptr, sub_len, search_start, search_len, enc);
if (pos < 0) return pos;
t = rb_enc_right_char_head(search_start, search_start+pos, str_ptr_end, enc);
if (t == search_start + pos) break;
search_len -= t - search_start;
if (search_len <= 0) return -1;
offset += t - search_start;
search_start = t;
}
return pos + offset;
}
#define rb_str_index(str, sub, offset) rb_strseq_index(str, sub, offset, 0)
static long
rb_strseq_index(VALUE str, VALUE sub, long offset, int in_byte)
{
const char *str_ptr, *str_ptr_end, *sub_ptr, *search_start;
long pos, str_len, sub_len, search_len;
const char *str_ptr, *str_ptr_end, *sub_ptr;
long str_len, sub_len;
int single_byte = single_byte_optimizable(str);
rb_encoding *enc;
......
if (sub_len == 0) return offset;
/* need proceed one character at a time */
search_start = str_ptr;
search_len = RSTRING_LEN(str) - offset;
for (;;) {
const char *t;
pos = rb_memsearch(sub_ptr, sub_len, search_start, search_len, enc);
if (pos < 0) return pos;
t = rb_enc_right_char_head(search_start, search_start+pos, str_ptr_end, enc);
if (t == search_start + pos) break;
search_len -= t - search_start;
if (search_len <= 0) return -1;
offset += t - search_start;
search_start = t;
}
return pos + offset;
return strseq_core(str_ptr, str_ptr_end, str_len, sub_ptr, sub_len, offset, enc);
}
......
return result;
}
/* Is s wrapped with '"'? */
static int
is_wrapped(const char *s, const char *s_end, long len, rb_encoding *enc)
{
unsigned int cbeg, cend;
const char *prev;
if (len < 2) return FALSE;
cbeg = rb_enc_mbc_to_codepoint(s, s_end, enc);
if (cbeg != '"') return FALSE;
prev = rb_enc_prev_char(s, s_end, s_end, enc);
cend = rb_enc_mbc_to_codepoint(prev, s_end, enc);
return cend == '"';
}
static const char *
unescape_ascii(unsigned int c)
{
switch (c) {
case 'n':
return "\n";
case 'r':
return "\r";
case 't':
return "\t";
case 'f':
return "\f";
case 'v':
return "\v";
case 'b':
return "\b";
case 'a':
return "\a";
case 'e':
return "\e";
default:
UNREACHABLE;
}
}
static int
undump_after_backslash(VALUE undumped, const char *s, const char *s_end, rb_encoding *enc)
{
unsigned int c, c2;
int n, n2, codelen;
size_t hexlen;
char buf[6];
c = rb_enc_codepoint_len(s, s_end, &n, enc);
switch (c) {
case '\\':
case '"':
rb_str_cat(undumped, s, n); /* cat itself */
n++;
break;
case 'n':
case 'r':
case 't':
case 'f':
case 'v':
case 'b':
case 'a':
case 'e':
rb_str_cat(undumped, unescape_ascii(c), n);
n++;
break;
case 'u':
if (s+1 >= s_end) {
rb_raise(rb_eArgError, "invalid Unicode escape");
}
c2 = rb_enc_codepoint_len(s+1, s_end, NULL, enc);
if (c2 == '{') { /* handle \u{...} form */
const char *hexstr = s + 2;
unsigned int hex;
static const char* const close_brace = "}";
long pos;
if (hexstr >= s_end) {
rb_raise(rb_eArgError, "unterminated Unicode escape");
}
/* find close brace */
pos = strseq_core(hexstr, s_end, s_end - hexstr, close_brace, 1, 0, enc);
if (pos < 0) {
rb_raise(rb_eArgError, "unterminated Unicode escape");
}
hex = ruby_scan_hex(hexstr, pos, &hexlen);
if (hexlen == 0 || hexlen > 6) {
rb_raise(rb_eArgError, "invalid Unicode escape");
}
if (hex > 0x10ffffU) {
rb_raise(rb_eArgError, "invalid Unicode codepoint (too large)");
}
if ((hex & 0xfffff800U) == 0xd800U) {
rb_raise(rb_eArgError, "invalid Unicode codepoint");
}
codelen = rb_enc_codelen(hex, enc);
rb_enc_mbcput(hex, buf, enc);
rb_str_cat(undumped, buf, codelen);
n += rb_strlen_lit("u{}") + hexlen;
}
else { /* handle \uXXXX form */
unsigned int hex = ruby_scan_hex(s+1, 4, &hexlen);
if (hexlen != 4) {
rb_raise(rb_eArgError, "invalid Unicode escape");
}
codelen = rb_enc_codelen(hex, enc);
rb_enc_mbcput(hex, buf, enc);
rb_str_cat(undumped, buf, codelen);
n += rb_strlen_lit("uXXXX");
}
break;
case 'x':
if (s+1 >= s_end) {
rb_raise(rb_eArgError, "invalid hex escape");
}
c2 = ruby_scan_hex(s+1, 2, &hexlen);
if (hexlen != 2) {
rb_raise(rb_eArgError, "invalid hex escape");
}
*buf = (char)c2;
rb_str_cat(undumped, buf, 1L);
n += rb_strlen_lit("xXX");
break;
case '#':
if (s+1 >= s_end) {
rb_str_cat(undumped, s, 1L); /* just '#' */
n++;
break;
}
n2 = rb_enc_mbclen(s+1, s_end, enc);
if (n2 == 1 && IS_EVSTR(s+1, s_end)) {
rb_str_cat(undumped, s, n);
n += n2;
}
break;
default:
rb_str_cat(undumped, "\\", 1L); /* keep backslash */
}
return n;
}
/*
* call-seq:
* str.undump -> new_str
*
* Produces unescaped version of +str+.
* See also String#dump because String#undump does inverse of String#dump.
*
* "\"hello \\n ''\"".undump #=> "hello \n ''"
*/
static VALUE
str_undump(VALUE str)
{
const char *s = RSTRING_PTR(str);
const char *s_end = RSTRING_END(str);
long len = RSTRING_LEN(str);
rb_encoding *enc = rb_enc_get(str);
int n;
unsigned int c;
VALUE undumped = rb_enc_str_new(s, 0L, enc);
rb_must_asciicompat(str);
if (is_wrapped(s, s_end, len, enc)) {
/* strip '"' at the begin and the end */
s++;
s_end--;
}
for (; s < s_end; s += n) {
c = rb_enc_codepoint_len(s, s_end, &n, enc);
if (c == '\\') {
if (s+1 >= s_end) {
rb_raise(rb_eArgError, "invalid escape");
}
n = undump_after_backslash(undumped, s+1, s_end, enc);
}
else {
rb_str_cat(undumped, s, n);
}
}
OBJ_INFECT(undumped, str);
return undumped;
}
static void
rb_str_check_dummy_enc(rb_encoding *enc)
......
rb_define_method(rb_cString, "to_str", rb_str_to_s, 0);
rb_define_method(rb_cString, "inspect", rb_str_inspect, 0);
rb_define_method(rb_cString, "dump", rb_str_dump, 0);
rb_define_method(rb_cString, "undump", str_undump, 0);
sym_ascii = ID2SYM(rb_intern("ascii"));
sym_turkic = ID2SYM(rb_intern("turkic"));
test/ruby/test_string.rb
assert_equal(S('"\\u{10ABCD}"'), b.dump)
end
def test_undump
a = S("Test") << 1 << 2 << 3 << 9 << 13 << 10
assert_equal(a, S('"Test\\x01\\x02\\x03\\t\\r\\n"').undump)
assert_equal(S("\u{7F}"), S('"\\x7F"').undump)
assert_equal(S("\u{AB}"), S('"\\u00AB"').undump)
assert_equal(S("\u{ABC}"), S('"\\u0ABC"').undump)
assert_equal(S("\uABCD"), S('"\\uABCD"').undump)
assert_equal(S("\u{ABCDE}"), S('"\\u{ABCDE}"').undump)
assert_equal(S("\u{10ABCD}"), S('"\\u{10ABCD}"').undump)
assert_equal(S("äöü"), S('"\u00E4\u00F6\u00FC"').undump)
assert_equal(S("äöü"), S('"\xC3\xA4\xC3\xB6\xC3\xBC"').undump)
end
def test_dup
for taint in [ false, true ]
for frozen in [ false, true ]
(2-2/5)