* string.c (str_gsub): should preserve last successful match

data.  [ruby-dev:35182]

git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@17447 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
This commit is contained in:
matz 2008-06-19 17:11:55 +00:00
parent e929f49e56
commit 5336cbdfd3
3 changed files with 22 additions and 19 deletions

View File

@ -1,3 +1,8 @@
Fri Jun 20 02:11:01 2008 Yukihiro Matsumoto <matz@ruby-lang.org>
* string.c (str_gsub): should preserve last successful match
data. [ruby-dev:35182]
Fri Jun 20 01:07:28 2008 Koichi Sasada <ko1@atdot.net> Fri Jun 20 01:07:28 2008 Koichi Sasada <ko1@atdot.net>
* KNOWNBUGS.rb, bootstraptest/pending.rb: move a bug (?) to pending. * KNOWNBUGS.rb, bootstraptest/pending.rb: move a bug (?) to pending.

View File

@ -392,6 +392,5 @@ assert_equal 'ok', %q{
f = Foo.new f = Foo.new
a_proc = give_it a_proc = give_it
p :call_it
f.call_it(&give_it()) f.call_it(&give_it())
}, '[ruby-core:15711]' }, '[ruby-core:15711]'

View File

@ -3084,8 +3084,7 @@ get_pat(VALUE pat, int quote)
static VALUE static VALUE
rb_str_sub_bang(int argc, VALUE *argv, VALUE str) rb_str_sub_bang(int argc, VALUE *argv, VALUE str)
{ {
VALUE pat, repl, match, hash = Qnil; VALUE pat, repl, hash = Qnil;
struct re_registers *regs;
int iter = 0; int iter = 0;
int tainted = 0; int tainted = 0;
long plen; long plen;
@ -3109,15 +3108,13 @@ rb_str_sub_bang(int argc, VALUE *argv, VALUE str)
if (rb_reg_search(pat, str, 0, 0) >= 0) { if (rb_reg_search(pat, str, 0, 0) >= 0) {
rb_encoding *enc; rb_encoding *enc;
int cr = ENC_CODERANGE(str); int cr = ENC_CODERANGE(str);
VALUE match = rb_backref_get();
match = rb_backref_get(); struct re_registers *regs = RMATCH_REGS(match);
regs = RMATCH_REGS(match);
if (iter || !NIL_P(hash)) { if (iter || !NIL_P(hash)) {
char *p = RSTRING_PTR(str); long len = RSTRING_LEN(str); char *p = RSTRING_PTR(str); long len = RSTRING_LEN(str);
if (iter) { if (iter) {
rb_match_busy(match);
repl = rb_obj_as_string(rb_yield(rb_reg_nth_match(0, match))); repl = rb_obj_as_string(rb_yield(rb_reg_nth_match(0, match)));
} }
else { else {
@ -3126,7 +3123,6 @@ rb_str_sub_bang(int argc, VALUE *argv, VALUE str)
} }
str_mod_check(str, p, len); str_mod_check(str, p, len);
str_frozen_check(str); str_frozen_check(str);
if (iter) rb_backref_set(match);
} }
else { else {
repl = rb_reg_regsub(repl, str, regs, pat); repl = rb_reg_regsub(repl, str, regs, pat);
@ -3217,7 +3213,8 @@ str_gsub(int argc, VALUE *argv, VALUE str, int bang)
VALUE pat, val, repl, match, dest, hash = Qnil; VALUE pat, val, repl, match, dest, hash = Qnil;
struct re_registers *regs; struct re_registers *regs;
long beg, n; long beg, n;
long offset, blen, slen, len; long beg0, end0;
long offset, blen, slen, len, last;
int iter = 0; int iter = 0;
char *sp, *cp; char *sp, *cp;
int tainted = 0; int tainted = 0;
@ -3241,13 +3238,14 @@ str_gsub(int argc, VALUE *argv, VALUE str, int bang)
} }
pat = get_pat(argv[0], 1); pat = get_pat(argv[0], 1);
offset=0; n=0;
beg = rb_reg_search(pat, str, 0, 0); beg = rb_reg_search(pat, str, 0, 0);
if (beg < 0) { if (beg < 0) {
if (bang) return Qnil; /* no match, no substitution */ if (bang) return Qnil; /* no match, no substitution */
return rb_str_dup(str); return rb_str_dup(str);
} }
offset = 0;
n = 0;
blen = RSTRING_LEN(str) + 30; /* len + margin */ blen = RSTRING_LEN(str) + 30; /* len + margin */
dest = rb_str_buf_new(blen); dest = rb_str_buf_new(blen);
sp = RSTRING_PTR(str); sp = RSTRING_PTR(str);
@ -3259,9 +3257,10 @@ str_gsub(int argc, VALUE *argv, VALUE str, int bang)
n++; n++;
match = rb_backref_get(); match = rb_backref_get();
regs = RMATCH_REGS(match); regs = RMATCH_REGS(match);
beg0 = BEG(0);
end0 = END(0);
if (iter || !NIL_P(hash)) { if (iter || !NIL_P(hash)) {
if (iter) { if (iter) {
rb_match_busy(match);
val = rb_obj_as_string(rb_yield(rb_reg_nth_match(0, match))); val = rb_obj_as_string(rb_yield(rb_reg_nth_match(0, match)));
} }
else { else {
@ -3273,7 +3272,6 @@ str_gsub(int argc, VALUE *argv, VALUE str, int bang)
if (val == dest) { /* paranoid check [ruby-dev:24827] */ if (val == dest) { /* paranoid check [ruby-dev:24827] */
rb_raise(rb_eRuntimeError, "block should not cheat"); rb_raise(rb_eRuntimeError, "block should not cheat");
} }
if (iter) rb_backref_set(match);
} }
else { else {
val = rb_reg_regsub(repl, str, regs, pat); val = rb_reg_regsub(repl, str, regs, pat);
@ -3288,16 +3286,17 @@ str_gsub(int argc, VALUE *argv, VALUE str, int bang)
rb_str_buf_append(dest, val); rb_str_buf_append(dest, val);
offset = END(0); last = offset;
if (BEG(0) == END(0)) { offset = end0;
if (beg0 == end0) {
/* /*
* Always consume at least one character of the input string * Always consume at least one character of the input string
* in order to prevent infinite loops. * in order to prevent infinite loops.
*/ */
if (RSTRING_LEN(str) <= END(0)) break; if (RSTRING_LEN(str) <= end0) break;
len = rb_enc_mbclen(RSTRING_PTR(str)+END(0), RSTRING_END(str), str_enc); len = rb_enc_mbclen(RSTRING_PTR(str)+end0, RSTRING_END(str), str_enc);
rb_enc_str_buf_cat(dest, RSTRING_PTR(str)+END(0), len, str_enc); rb_enc_str_buf_cat(dest, RSTRING_PTR(str)+end0, len, str_enc);
offset = END(0) + len; offset = end0 + len;
} }
cp = RSTRING_PTR(str) + offset; cp = RSTRING_PTR(str) + offset;
if (offset > RSTRING_LEN(str)) break; if (offset > RSTRING_LEN(str)) break;
@ -3306,7 +3305,7 @@ str_gsub(int argc, VALUE *argv, VALUE str, int bang)
if (RSTRING_LEN(str) > offset) { if (RSTRING_LEN(str) > offset) {
rb_enc_str_buf_cat(dest, cp, RSTRING_LEN(str) - offset, str_enc); rb_enc_str_buf_cat(dest, cp, RSTRING_LEN(str) - offset, str_enc);
} }
rb_backref_set(match); rb_reg_search(pat, str, last, 0);
if (bang) { if (bang) {
rb_str_shared_replace(str, dest); rb_str_shared_replace(str, dest);
} }