A Range
represents an interval—a set of values with a
beginning and an end. Ranges may be constructed using the
s..
e and
s...
e literals, or with ::new. Ranges constructed using
..
run from the beginning to the end inclusively. Those
created using ...
exclude the end value. When used as an
iterator, ranges return each value in the sequence.
(-1..-5).to_a #=> [] (-5..-1).to_a #=> [-5, -4, -3, -2, -1] ('a'..'e').to_a #=> ["a", "b", "c", "d", "e"] ('a'...'e').to_a #=> ["a", "b", "c", "d"]
An “endless range” represents a semi-infinite range. Literal notation for an endless range is:
(1..) # or similarly (1...)
Which is equivalent to
(1..nil) # or similarly (1...nil) Range.new(1, nil) # or Range.new(1, nil, true)
Endless ranges are useful, for example, for idiomatic slicing of arrays:
[1, 2, 3, 4, 5][2...] # => [3, 4, 5]
Some implementation details:
end
of endless range is nil
;
each
of endless range enumerates infinite sequence (may be
useful in combination with Enumerable#take_while or
similar methods);
(1..)
and (1...)
are not equal, although
technically representing the same sequence.
Ranges can be constructed using any objects that can be compared using the
<=>
operator. Methods that treat the range as a sequence
(#each and methods inherited from Enumerable)
expect the begin object to implement a succ
method to return
the next object in sequence. The step and include? methods require the
begin object to implement succ
or to be numeric.
In the Xs
class below both <=>
and
succ
are implemented so Xs
can be used to
construct ranges. Note that the Comparable
module is included so the ==
method is defined in terms of
<=>
.
class Xs # represent a string of 'x's include Comparable attr :length def initialize(n) @length = n end def succ Xs.new(@length + 1) end def <=>(other) @length <=> other.length end def to_s sprintf "%2d #{inspect}", @length end def inspect 'x' * @length end end
An example of using Xs
to construct a range:
r = Xs.new(3)..Xs.new(6) #=> xxx..xxxxxx r.to_a #=> [xxx, xxxx, xxxxx, xxxxxx] r.member?(Xs.new(5)) #=> true
Constructs a range using the given begin
and end
.
If the exclude_end
parameter is omitted or is
false
, the range will include the end object; otherwise, it
will be excluded.
static VALUE range_initialize(int argc, VALUE *argv, VALUE range) { VALUE beg, end, flags; rb_scan_args(argc, argv, "21", &beg, &end, &flags); range_modify(range); range_init(range, beg, end, RBOOL(RTEST(flags))); return Qnil; }
Iterates over the range, passing each n
th element to the
block. If begin and end are numeric, n
is added for each
iteration. Otherwise step
invokes succ
to iterate
through range elements.
If no block is given, an enumerator is returned instead. Especially, the enumerator is an Enumerator::ArithmeticSequence if begin and end of the range are numeric.
range = Xs.new(1)..Xs.new(10) range.step(2) {|x| puts x} puts range.step(3) {|x| puts x}
produces:
1 x 3 xxx 5 xxxxx 7 xxxxxxx 9 xxxxxxxxx 1 x 4 xxxx 7 xxxxxxx 10 xxxxxxxxxx
See Range for the definition of class Xs.
static VALUE range_percent_step(VALUE range, VALUE step) { return range_step(1, &step, range); }
Returns true
only if obj
is a Range, has equivalent begin and end items (by
comparing them with ==
), and has the same exclude_end? setting as the
range.
(0..2) == (0..2) #=> true (0..2) == Range.new(0,2) #=> true (0..2) == (0...2) #=> false
static VALUE range_eq(VALUE range, VALUE obj) { if (range == obj) return Qtrue; if (!rb_obj_is_kind_of(obj, rb_cRange)) return Qfalse; return rb_exec_recursive_paired(recursive_equal, range, obj, obj); }
Returns true
if obj
is an element of the range,
false
otherwise. Conveniently, ===
is the
comparison operator used by case
statements.
case 79 when 1..50 then print "low\n" when 51..75 then print "medium\n" when 76..100 then print "high\n" end
produces:
high
static VALUE range_eqq(VALUE range, VALUE val) { VALUE ret = range_include_internal(range, val); if (ret != Qundef) return ret; return r_cover_p(range, RANGE_BEG(range), RANGE_END(range), val); }
Returns the object that defines the beginning of the range.
(1..10).begin #=> 1
static VALUE range_begin(VALUE range) { return RANGE_BEG(range); }
By using binary search, finds a value in range which meets the given condition in O(log n) where n is the size of the range.
You can use this method in two use cases: a find-minimum mode and a find-any mode. In either case, the elements of the range must be monotone (or sorted) with respect to the block.
In find-minimum mode (this is a good choice for typical use case), the block must return true or false, and there must be a value x so that:
the block returns false for any value which is less than x, and
the block returns true for any value which is greater than or equal to x.
If x is within the range, this method returns the value x. Otherwise, it returns nil.
ary = [0, 4, 7, 10, 12] (0...ary.size).bsearch {|i| ary[i] >= 4 } #=> 1 (0...ary.size).bsearch {|i| ary[i] >= 6 } #=> 2 (0...ary.size).bsearch {|i| ary[i] >= 8 } #=> 3 (0...ary.size).bsearch {|i| ary[i] >= 100 } #=> nil (0.0...Float::INFINITY).bsearch {|x| Math.log(x) >= 0 } #=> 1.0
In find-any mode (this behaves like libc's bsearch(3)), the block must return a number, and there must be two values x and y (x <= y) so that:
the block returns a positive number for v if v < x,
the block returns zero for v if x <= v < y, and
the block returns a negative number for v if y <= v.
This method returns any value which is within the intersection of the given range and x…y (if any). If there is no value that satisfies the condition, it returns nil.
ary = [0, 100, 100, 100, 200] (0..4).bsearch {|i| 100 - ary[i] } #=> 1, 2 or 3 (0..4).bsearch {|i| 300 - ary[i] } #=> nil (0..4).bsearch {|i| 50 - ary[i] } #=> nil
You must not mix the two modes at a time; the block must always return either true/false, or always return a number. It is undefined which value is actually picked up at each iteration.
static VALUE range_bsearch(VALUE range) { VALUE beg, end, satisfied = Qnil; int smaller; /* Implementation notes: * Floats are handled by mapping them to 64 bits integers. * Apart from sign issues, floats and their 64 bits integer have the * same order, assuming they are represented as exponent followed * by the mantissa. This is true with or without implicit bit. * * Finding the average of two ints needs to be careful about * potential overflow (since float to long can use 64 bits) * as well as the fact that -1/2 can be 0 or -1 in C89. * * Note that -0.0 is mapped to the same int as 0.0 as we don't want * (-1...0.0).bsearch to yield -0.0. */ #define BSEARCH(conv) \ do { \ RETURN_ENUMERATOR(range, 0, 0); \ if (EXCL(range)) high--; \ org_high = high; \ while (low < high) { \ mid = ((high < 0) == (low < 0)) ? low + ((high - low) / 2) \ : (low < -high) ? -((-1 - low - high)/2 + 1) : (low + high) / 2; \ BSEARCH_CHECK(conv(mid)); \ if (smaller) { \ high = mid; \ } \ else { \ low = mid + 1; \ } \ } \ if (low == org_high) { \ BSEARCH_CHECK(conv(low)); \ if (!smaller) return Qnil; \ } \ return satisfied; \ } while (0) beg = RANGE_BEG(range); end = RANGE_END(range); if (FIXNUM_P(beg) && FIXNUM_P(end)) { long low = FIX2LONG(beg); long high = FIX2LONG(end); long mid, org_high; BSEARCH(INT2FIX); } #if SIZEOF_DOUBLE == 8 && defined(HAVE_INT64_T) else if (RB_TYPE_P(beg, T_FLOAT) || RB_TYPE_P(end, T_FLOAT)) { int64_t low = double_as_int64(RFLOAT_VALUE(rb_Float(beg))); int64_t high = double_as_int64(NIL_P(end) ? HUGE_VAL : RFLOAT_VALUE(rb_Float(end))); int64_t mid, org_high; BSEARCH(int64_as_double_to_num); } #endif else if (is_integer_p(beg) && is_integer_p(end)) { RETURN_ENUMERATOR(range, 0, 0); return bsearch_integer_range(beg, end, EXCL(range)); } else if (is_integer_p(beg) && NIL_P(end)) { VALUE diff = LONG2FIX(1); RETURN_ENUMERATOR(range, 0, 0); while (1) { VALUE mid = rb_funcall(beg, '+', 1, diff); BSEARCH_CHECK(mid); if (smaller) { return bsearch_integer_range(beg, mid, 0); } diff = rb_funcall(diff, '*', 1, LONG2FIX(2)); } } else { rb_raise(rb_eTypeError, "can't do binary search for %s", rb_obj_classname(beg)); } return range; }
Returns true
if obj
is between the begin and end
of the range.
This tests begin <= obj <= end
when exclude_end? is
false
and begin <= obj < end
when exclude_end? is
true
.
If called with a Range argument, returns
true
when the given range is covered by the receiver, by
comparing the begin and end values. If the argument can be treated as a
sequence, this method treats it that way. In the specific case of
(a..b).cover?(c...d)
with a <= c && b <
d
, the end of the sequence must be calculated, which may exhibit
poor performance if c
is non-numeric. Returns
false
if the begin value of the range is larger than the end
value.
("a".."z").cover?("c") #=> true ("a".."z").cover?("5") #=> false ("a".."z").cover?("cc") #=> true (1..5).cover?(2..3) #=> true (1..5).cover?(0..6) #=> false (1..5).cover?(1...6) #=> true
static VALUE range_cover(VALUE range, VALUE val) { VALUE beg, end; beg = RANGE_BEG(range); end = RANGE_END(range); if (rb_obj_is_kind_of(val, rb_cRange)) { return RBOOL(r_cover_range_p(range, beg, end, val)); } return r_cover_p(range, beg, end, val); }
Iterates over the elements of range, passing each in turn to the block.
The each
method can only be used if the begin object of the
range supports the succ
method. A TypeError is raised if the object does not have
succ
method defined (like Float).
If no block is given, an enumerator is returned instead.
(10..15).each {|n| print n, ' ' } # prints: 10 11 12 13 14 15 (2.5..5).each {|n| print n, ' ' } # raises: TypeError: can't iterate from Float
static VALUE range_each(VALUE range) { VALUE beg, end; long i, lim; RETURN_SIZED_ENUMERATOR(range, 0, 0, range_enum_size); beg = RANGE_BEG(range); end = RANGE_END(range); if (FIXNUM_P(beg) && NIL_P(end)) { fixnum_endless: i = FIX2LONG(beg); while (FIXABLE(i)) { rb_yield(LONG2FIX(i++)); } beg = LONG2NUM(i); bignum_endless: for (;; beg = rb_big_plus(beg, INT2FIX(1))) rb_yield(beg); } else if (FIXNUM_P(beg) && FIXNUM_P(end)) { /* fixnums are special */ fixnum_loop: lim = FIX2LONG(end); if (!EXCL(range)) lim += 1; for (i = FIX2LONG(beg); i < lim; i++) { rb_yield(LONG2FIX(i)); } } else if (RB_INTEGER_TYPE_P(beg) && (NIL_P(end) || RB_INTEGER_TYPE_P(end))) { if (SPECIAL_CONST_P(end) || RBIGNUM_POSITIVE_P(end)) { /* end >= FIXNUM_MIN */ if (!FIXNUM_P(beg)) { if (RBIGNUM_NEGATIVE_P(beg)) { do { rb_yield(beg); } while (!FIXNUM_P(beg = rb_big_plus(beg, INT2FIX(1)))); if (NIL_P(end)) goto fixnum_endless; if (FIXNUM_P(end)) goto fixnum_loop; } else { if (NIL_P(end)) goto bignum_endless; if (FIXNUM_P(end)) return range; } } if (FIXNUM_P(beg)) { i = FIX2LONG(beg); do { rb_yield(LONG2FIX(i)); } while (POSFIXABLE(++i)); beg = LONG2NUM(i); } ASSUME(!FIXNUM_P(beg)); ASSUME(!SPECIAL_CONST_P(end)); } if (!FIXNUM_P(beg) && RBIGNUM_SIGN(beg) == RBIGNUM_SIGN(end)) { if (EXCL(range)) { while (rb_big_cmp(beg, end) == INT2FIX(-1)) { rb_yield(beg); beg = rb_big_plus(beg, INT2FIX(1)); } } else { VALUE c; while ((c = rb_big_cmp(beg, end)) != INT2FIX(1)) { rb_yield(beg); if (c == INT2FIX(0)) break; beg = rb_big_plus(beg, INT2FIX(1)); } } } } else if (SYMBOL_P(beg) && (NIL_P(end) || SYMBOL_P(end))) { /* symbols are special */ beg = rb_sym2str(beg); if (NIL_P(end)) { rb_str_upto_endless_each(beg, sym_each_i, 0); } else { rb_str_upto_each(beg, rb_sym2str(end), EXCL(range), sym_each_i, 0); } } else { VALUE tmp = rb_check_string_type(beg); if (!NIL_P(tmp)) { if (!NIL_P(end)) { rb_str_upto_each(tmp, end, EXCL(range), each_i, 0); } else { rb_str_upto_endless_each(tmp, each_i, 0); } } else { if (!discrete_object_p(beg)) { rb_raise(rb_eTypeError, "can't iterate from %s", rb_obj_classname(beg)); } if (!NIL_P(end)) range_each_func(range, each_i, 0); else for (;; beg = rb_funcallv(beg, id_succ, 0, 0)) rb_yield(beg); } } return range; }
Returns the object that defines the end of the range.
(1..10).end #=> 10 (1...10).end #=> 10
static VALUE range_end(VALUE range) { return RANGE_END(range); }
Returns an array containing the items in the range.
(1..7).to_a #=> [1, 2, 3, 4, 5, 6, 7] (1..).to_a #=> RangeError: cannot convert endless range to an array
static VALUE range_to_a(VALUE range) { if (NIL_P(RANGE_END(range))) { rb_raise(rb_eRangeError, "cannot convert endless range to an array"); } return rb_call_super(0, 0); }
Returns true
only if obj
is a Range, has equivalent begin and end items (by
comparing them with eql?
), and has the same exclude_end? setting as the
range.
(0..2).eql?(0..2) #=> true (0..2).eql?(Range.new(0,2)) #=> true (0..2).eql?(0...2) #=> false
static VALUE range_eql(VALUE range, VALUE obj) { if (range == obj) return Qtrue; if (!rb_obj_is_kind_of(obj, rb_cRange)) return Qfalse; return rb_exec_recursive_paired(recursive_eql, range, obj, obj); }
Returns true
if the range excludes its end value.
(1..5).exclude_end? #=> false (1...5).exclude_end? #=> true
static VALUE range_exclude_end_p(VALUE range) { return EXCL(range) ? Qtrue : Qfalse; }
Returns the first object in the range, or an array of the first
n
elements.
(10..20).first #=> 10 (10..20).first(3) #=> [10, 11, 12]
static VALUE range_first(int argc, VALUE *argv, VALUE range) { VALUE n, ary[2]; if (argc == 0) return RANGE_BEG(range); rb_scan_args(argc, argv, "1", &n); ary[0] = n; ary[1] = rb_ary_new2(NUM2LONG(n)); rb_block_call(range, idEach, 0, 0, first_i, (VALUE)ary); return ary[1]; }
Compute a hash-code for this range. Two ranges with equal begin and end
points (using eql?
), and the same exclude_end? value will
generate the same hash-code.
See also Object#hash.
static VALUE range_hash(VALUE range) { st_index_t hash = EXCL(range); VALUE v; hash = rb_hash_start(hash); v = rb_hash(RANGE_BEG(range)); hash = rb_hash_uint(hash, NUM2LONG(v)); v = rb_hash(RANGE_END(range)); hash = rb_hash_uint(hash, NUM2LONG(v)); hash = rb_hash_uint(hash, EXCL(range) << 24); hash = rb_hash_end(hash); return LONG2FIX(hash); }
Returns true
if obj
is an element of the range,
false
otherwise. If begin and end are numeric, comparison is
done according to the magnitude of the values.
("a".."z").include?("g") #=> true ("a".."z").include?("A") #=> false ("a".."z").include?("cc") #=> false
static VALUE range_include(VALUE range, VALUE val) { VALUE ret = range_include_internal(range, val); if (ret != Qundef) return ret; return rb_call_super(1, &val); }
Convert this range object to a printable form (using inspect
to convert the begin and end objects).
static VALUE range_inspect(VALUE range) { return rb_exec_recursive(inspect_range, range, 0); }
Returns the last object in the range, or an array of the last
n
elements.
Note that with no arguments last
will return the object that
defines the end of the range even if exclude_end? is
true
.
(10..20).last #=> 20 (10...20).last #=> 20 (10..20).last(3) #=> [18, 19, 20] (10...20).last(3) #=> [17, 18, 19]
static VALUE range_last(int argc, VALUE *argv, VALUE range) { if (NIL_P(RANGE_END(range))) { rb_raise(rb_eRangeError, "cannot get the last element of endless range"); } if (argc == 0) return RANGE_END(range); return rb_ary_last(argc, argv, rb_Array(range)); }
Returns the maximum value in the range. Returns nil
if the
begin value of the range larger than the end value. Returns
nil
if the begin value of an exclusive range is equal to the
end value.
Can be given an optional block to override the default comparison method
a <=> b
.
(10..20).max #=> 20
static VALUE range_max(int argc, VALUE *argv, VALUE range) { VALUE e = RANGE_END(range); int nm = FIXNUM_P(e) || rb_obj_is_kind_of(e, rb_cNumeric); if (NIL_P(RANGE_END(range))) { rb_raise(rb_eRangeError, "cannot get the maximum of endless range"); } if (rb_block_given_p() || (EXCL(range) && !nm) || argc) { return rb_call_super(argc, argv); } else { struct cmp_opt_data cmp_opt = { 0, 0 }; VALUE b = RANGE_BEG(range); int c = OPTIMIZED_CMP(b, e, cmp_opt); if (c > 0) return Qnil; if (EXCL(range)) { if (!RB_INTEGER_TYPE_P(e)) { rb_raise(rb_eTypeError, "cannot exclude non Integer end value"); } if (c == 0) return Qnil; if (!RB_INTEGER_TYPE_P(b)) { rb_raise(rb_eTypeError, "cannot exclude end value with non Integer begin value"); } if (FIXNUM_P(e)) { return LONG2NUM(FIX2LONG(e) - 1); } return rb_funcall(e, '-', 1, INT2FIX(1)); } return e; } }
Returns true
if obj
is an element of the range,
false
otherwise. If begin and end are numeric, comparison is
done according to the magnitude of the values.
("a".."z").include?("g") #=> true ("a".."z").include?("A") #=> false ("a".."z").include?("cc") #=> false
static VALUE range_include(VALUE range, VALUE val) { VALUE ret = range_include_internal(range, val); if (ret != Qundef) return ret; return rb_call_super(1, &val); }
Returns the minimum value in the range. Returns nil
if the
begin value of the range is larger than the end value. Returns
nil
if the begin value of an exclusive range is equal to the
end value.
Can be given an optional block to override the default comparison method
a <=> b
.
(10..20).min #=> 10
static VALUE range_min(int argc, VALUE *argv, VALUE range) { if (rb_block_given_p()) { if (NIL_P(RANGE_END(range))) { rb_raise(rb_eRangeError, "cannot get the minimum of endless range with custom comparison method"); } return rb_call_super(argc, argv); } else if (argc != 0) { return range_first(argc, argv, range); } else { struct cmp_opt_data cmp_opt = { 0, 0 }; VALUE b = RANGE_BEG(range); VALUE e = RANGE_END(range); int c = NIL_P(e) ? -1 : OPTIMIZED_CMP(b, e, cmp_opt); if (c > 0 || (c == 0 && EXCL(range))) return Qnil; return b; } }
Returns the number of elements in the range. Both the begin and the end of the Range must be Numeric, otherwise nil is returned.
(10..20).size #=> 11 ('a'..'z').size #=> nil (-Float::INFINITY..Float::INFINITY).size #=> Infinity
static VALUE range_size(VALUE range) { VALUE b = RANGE_BEG(range), e = RANGE_END(range); if (rb_obj_is_kind_of(b, rb_cNumeric)) { if (rb_obj_is_kind_of(e, rb_cNumeric)) { return ruby_num_interval_step_size(b, e, INT2FIX(1), EXCL(range)); } if (NIL_P(e)) { return DBL2NUM(HUGE_VAL); } } return Qnil; }
%
Iterates over the range, passing each n
th element to the
block. If begin and end are numeric, n
is added for each
iteration. Otherwise step
invokes succ
to iterate
through range elements.
If no block is given, an enumerator is returned instead. Especially, the enumerator is an Enumerator::ArithmeticSequence if begin and end of the range are numeric.
range = Xs.new(1)..Xs.new(10) range.step(2) {|x| puts x} puts range.step(3) {|x| puts x}
produces:
1 x 3 xxx 5 xxxxx 7 xxxxxxx 9 xxxxxxxxx 1 x 4 xxxx 7 xxxxxxx 10 xxxxxxxxxx
See Range for the definition of class Xs.
static VALUE range_step(int argc, VALUE *argv, VALUE range) { VALUE b, e, step, tmp; b = RANGE_BEG(range); e = RANGE_END(range); step = (!rb_check_arity(argc, 0, 1) ? INT2FIX(1) : argv[0]); if (!rb_block_given_p()) { if (rb_obj_is_kind_of(b, rb_cNumeric) && (NIL_P(e) || rb_obj_is_kind_of(e, rb_cNumeric))) { return rb_arith_seq_new(range, ID2SYM(rb_frame_this_func()), argc, argv, range_step_size, b, e, step, EXCL(range)); } RETURN_SIZED_ENUMERATOR(range, argc, argv, range_step_size); } step = check_step_domain(step); if (FIXNUM_P(b) && NIL_P(e) && FIXNUM_P(step)) { long i = FIX2LONG(b), unit = FIX2LONG(step); do { rb_yield(LONG2FIX(i)); i += unit; /* FIXABLE+FIXABLE never overflow */ } while (FIXABLE(i)); b = LONG2NUM(i); for (;; b = rb_big_plus(b, step)) rb_yield(b); } else if (FIXNUM_P(b) && FIXNUM_P(e) && FIXNUM_P(step)) { /* fixnums are special */ long end = FIX2LONG(e); long i, unit = FIX2LONG(step); if (!EXCL(range)) end += 1; i = FIX2LONG(b); while (i < end) { rb_yield(LONG2NUM(i)); if (i + unit < i) break; i += unit; } } else if (SYMBOL_P(b) && (NIL_P(e) || SYMBOL_P(e))) { /* symbols are special */ VALUE iter[2]; iter[0] = INT2FIX(1); iter[1] = step; b = rb_sym2str(b); if (NIL_P(e)) { rb_str_upto_endless_each(b, sym_step_i, (VALUE)iter); } else { rb_str_upto_each(b, rb_sym2str(e), EXCL(range), sym_step_i, (VALUE)iter); } } else if (ruby_float_step(b, e, step, EXCL(range), TRUE)) { /* done */ } else if (rb_obj_is_kind_of(b, rb_cNumeric) || !NIL_P(rb_check_to_integer(b, "to_int")) || !NIL_P(rb_check_to_integer(e, "to_int"))) { ID op = EXCL(range) ? '<' : idLE; VALUE v = b; int i = 0; while (NIL_P(e) || RTEST(rb_funcall(v, op, 1, e))) { rb_yield(v); i++; v = rb_funcall(b, '+', 1, rb_funcall(INT2NUM(i), '*', 1, step)); } } else { tmp = rb_check_string_type(b); if (!NIL_P(tmp)) { VALUE iter[2]; b = tmp; iter[0] = INT2FIX(1); iter[1] = step; if (NIL_P(e)) { rb_str_upto_endless_each(b, step_i, (VALUE)iter); } else { rb_str_upto_each(b, e, EXCL(range), step_i, (VALUE)iter); } } else { VALUE args[2]; if (!discrete_object_p(b)) { rb_raise(rb_eTypeError, "can't iterate from %s", rb_obj_classname(b)); } args[0] = INT2FIX(1); args[1] = step; range_each_func(range, step_i, (VALUE)args); } } return range; }
Returns an array containing the items in the range.
(1..7).to_a #=> [1, 2, 3, 4, 5, 6, 7] (1..).to_a #=> RangeError: cannot convert endless range to an array
static VALUE range_to_a(VALUE range) { if (NIL_P(RANGE_END(range))) { rb_raise(rb_eRangeError, "cannot convert endless range to an array"); } return rb_call_super(0, 0); }
Convert this range object to a printable form (using to_s to convert the begin and end objects).
static VALUE range_to_s(VALUE range) { VALUE str, str2; str = rb_obj_as_string(RANGE_BEG(range)); str2 = rb_obj_as_string(RANGE_END(range)); str = rb_str_dup(str); rb_str_cat(str, "...", EXCL(range) ? 3 : 2); rb_str_append(str, str2); OBJ_INFECT(str, range); return str; }