aboutsummaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
Diffstat
-rw-r--r--ChangeLog5
-rw-r--r--include/ruby/encoding.h2
-rw-r--r--string.c126
-rw-r--r--version.h6
4 files changed, 109 insertions, 30 deletions
diff --git a/ChangeLog b/ChangeLog
index 00d7d2842c..878487e2d7 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,8 @@
+Thu Sep 6 12:42:10 2007 Nobuyoshi Nakada <nobu@ruby-lang.org>
+
+ * string.c (rb_str_succ, rb_str_chop_bang, rb_str_chop): m17n support.
+ [ruby-dev:31734]
+
Wed Sep 5 22:02:27 2007 Yukihiro Matsumoto <matz@ruby-lang.org>
* array.c (rb_ary_subseq): need integer overflow check.
diff --git a/include/ruby/encoding.h b/include/ruby/encoding.h
index 805209d15a..e807e8a547 100644
--- a/include/ruby/encoding.h
+++ b/include/ruby/encoding.h
@@ -62,7 +62,7 @@ int rb_enc_codelen(int, rb_encoding*);
#define rb_enc_codepoint(p,e,enc) ONIGENC_MBC_TO_CODE(enc,(UChar*)p,(UChar*)e)
/* ptr, ptr, encoding -> prev_char */
-#define rb_enc_prev_char(s,p,enc) onigenc_get_prev_char_head(enc,(UChar*)s,(UChar*)p)
+#define rb_enc_prev_char(s,p,enc) (char *)onigenc_get_prev_char_head(enc,(UChar*)s,(UChar*)p)
#define rb_enc_isascii(c,enc) ONIGENC_IS_CODE_ASCII(c)
#define rb_enc_isalpha(c,enc) ONIGENC_IS_CODE_ALPHA(enc,c)
diff --git a/string.c b/string.c
index 88c5bf4706..2618045ee3 100644
--- a/string.c
+++ b/string.c
@@ -1584,6 +1584,52 @@ succ_char(char *s)
return 0;
}
+/*
+ overwrite +s+ by succeeding letter of +c+ in +enc+ and returns
+ carried-out letter. assuming each ranges are successive, and mbclen
+ never change in each ranges.
+ */
+static int
+enc_succ_char(unsigned int c, char *s, rb_encoding *enc)
+{
+ unsigned int cs;
+
+ /* numerics */
+ if (rb_enc_isdigit(c, enc)) {
+ cs = c++;
+ if (rb_enc_isdigit(c, enc)) {
+ rb_enc_mbcput(c, s, enc);
+ return 0;
+ }
+ do c = cs--; while (rb_enc_isdigit(cs, enc));
+ rb_enc_mbcput(c, s, enc);
+ return ++c;
+ }
+ /* small alphabets */
+ if (rb_enc_islower(c, enc)) {
+ cs = c++;
+ if (rb_enc_islower(c, enc)) {
+ rb_enc_mbcput(c, s, enc);
+ return 0;
+ }
+ do c = cs--; while (rb_enc_islower(cs, enc));
+ rb_enc_mbcput(c, s, enc);
+ return c;
+ }
+ /* capital alphabets */
+ if (rb_enc_isupper(c, enc)) {
+ cs = c++;
+ if (rb_enc_isupper(c, enc)) {
+ rb_enc_mbcput(c, s, enc);
+ return 0;
+ }
+ do c = cs--; while (rb_enc_isupper(cs, enc));
+ rb_enc_mbcput(c, s, enc);
+ return c;
+ }
+ return -1;
+}
+
/*
* call-seq:
@@ -1617,38 +1663,51 @@ rb_str_succ(VALUE orig)
VALUE str;
char *sbeg, *s, *e;
int c = -1;
- long n = 0;
+ long n = 0, o = 0, l;
+ char carry[ONIGENC_CODE_TO_MBC_MAXLEN];
str = rb_str_new5(orig, RSTRING_PTR(orig), RSTRING_LEN(orig));
+ rb_enc_copy(str, orig);
OBJ_INFECT(str, orig);
if (RSTRING_LEN(str) == 0) return str;
enc = rb_enc_get(orig);
- sbeg = RSTRING_PTR(str); s = sbeg + RSTRING_LEN(str) - 1;
- e = RSTRING_END(str);
+ sbeg = RSTRING_PTR(str);
+ s = e = sbeg + RSTRING_LEN(str);
- while (sbeg <= s) {
+ while ((s = rb_enc_prev_char(sbeg, s, enc)) != 0) {
unsigned int cc = rb_enc_codepoint(s, e, enc);
if (rb_enc_isalnum(cc, enc)) {
- if ((c = succ_char(s)) == 0) break;
+ if (isascii(cc)) {
+ if ((c = succ_char(s)) == 0) break;
+ }
+ else {
+ if ((c = enc_succ_char(cc, s, enc)) == 0) break;
+ }
n = s - sbeg;
}
- s--;
}
if (c == -1) { /* str contains no alnum */
- sbeg = RSTRING_PTR(str); s = sbeg + RSTRING_LEN(str) - 1;
c = '\001';
- while (sbeg <= s) {
- if ((*s += 1) != 0) break;
- s--;
+ s = e;
+ while ((s = rb_enc_prev_char(sbeg, e, enc)) != 0) {
+ unsigned int cc = rb_enc_codepoint(s, e, enc) + 1;
+ l = rb_enc_mbcput(cc, carry, enc);
+ if (l > 0) {
+ if (l == (o = e - s)) goto overlay;
+ n = s - sbeg;
+ goto insert;
+ }
}
}
- if (s < sbeg) {
- RESIZE_CAPA(str, RSTRING_LEN(str) + 1);
+ if (!s && (l = rb_enc_mbcput(c, carry, enc)) > 0) {
+ insert:
+ RESIZE_CAPA(str, RSTRING_LEN(str) + l - o);
s = RSTRING_PTR(str) + n;
- memmove(s+1, s, RSTRING_LEN(str) - n);
- *s = c;
- STR_SET_LEN(str, RSTRING_LEN(str) + 1);
+ memmove(s + l, s + o, RSTRING_LEN(str) - n - o);
+ overlay:
+ memmove(s, carry, l);
+ STR_SET_LEN(str, RSTRING_LEN(str) + l - o);
RSTRING_PTR(str)[RSTRING_LEN(str)] = '\0';
}
@@ -4040,6 +4099,24 @@ rb_str_each_char(VALUE str)
return str;
}
+static long
+chopped_length(VALUE str)
+{
+ rb_encoding *enc = rb_enc_get(str);
+ const char *p, *p2, *beg, *end;
+
+ beg = RSTRING_PTR(str);
+ end = beg + RSTRING_LEN(str);
+ if (beg > end) return 0;
+ p = rb_enc_prev_char(beg, end, enc);
+ if (!p) return 0;
+ if (p > beg && rb_enc_codepoint(p, end, enc) == '\n') {
+ p2 = rb_enc_prev_char(beg, p, enc);
+ if (p2 && rb_enc_codepoint(p2, end, enc) == '\r') p = p2;
+ }
+ return p - beg;
+}
+
/*
* call-seq:
* str.chop! => str or nil
@@ -4053,15 +4130,11 @@ static VALUE
rb_str_chop_bang(VALUE str)
{
if (RSTRING_LEN(str) > 0) {
+ long len;
rb_str_modify(str);
- STR_DEC_LEN(str);
- if (RSTRING_PTR(str)[RSTRING_LEN(str)] == '\n') {
- if (RSTRING_LEN(str) > 0 &&
- RSTRING_PTR(str)[RSTRING_LEN(str)-1] == '\r') {
- STR_DEC_LEN(str);
- }
- }
- RSTRING_PTR(str)[RSTRING_LEN(str)] = '\0';
+ len = chopped_length(str);
+ STR_SET_LEN(str, len);
+ RSTRING_PTR(str)[len] = '\0';
return str;
}
return Qnil;
@@ -4088,9 +4161,10 @@ rb_str_chop_bang(VALUE str)
static VALUE
rb_str_chop(VALUE str)
{
- str = rb_str_dup(str);
- rb_str_chop_bang(str);
- return str;
+ VALUE str2 = rb_str_new5(str, RSTRING_PTR(str), chopped_length(str));
+ rb_enc_copy(str2, str);
+ OBJ_INFECT(str2, str);
+ return str2;
}
diff --git a/version.h b/version.h
index c49a73f30b..389c0d4bc0 100644
--- a/version.h
+++ b/version.h
@@ -1,7 +1,7 @@
#define RUBY_VERSION "1.9.0"
-#define RUBY_RELEASE_DATE "2007-09-05"
+#define RUBY_RELEASE_DATE "2007-09-06"
#define RUBY_VERSION_CODE 190
-#define RUBY_RELEASE_CODE 20070905
+#define RUBY_RELEASE_CODE 20070906
#define RUBY_PATCHLEVEL 0
#define RUBY_VERSION_MAJOR 1
@@ -9,7 +9,7 @@
#define RUBY_VERSION_TEENY 0
#define RUBY_RELEASE_YEAR 2007
#define RUBY_RELEASE_MONTH 9
-#define RUBY_RELEASE_DAY 5
+#define RUBY_RELEASE_DAY 6
#ifdef RUBY_EXTERN
RUBY_EXTERN const char ruby_version[];
generated by cgit v1.2.3 (git 2.39.1) at 2024-05-27 00:39:39 +0000