Feature #7883: Add Regex#to_proc

• Status: Assigned
• Assignee: matz (Yukihiro Matsumoto)
• Target version: -

[rklemme (Robert Klemme)]

Just a small addition to the standard library:

class Regexp
  def to_proc; lambda {|s| self =~ s} end
end

With that one can use a Regexp everywhere a Proc is used as filtering criteria saving a bit of typing. While we have Enumerable#grep already there may be other cases where you want to do something like

irb(main):008:0> %w{foo bar baz}.select &/\Ab/
=> ["bar", "baz"]
irb(main):009:0> %w{foo bar baz}.reject &/\Ab/
=> ["foo"]
irb(main):010:0> %w{foo bar baz}.find &/\Ab/
=> "bar"

Note: line 9 and 10 are not possible with Enumerable#grep AFAIK.

I see low risk of breaking something.

[nobu (Nobuyoshi Nakada)]

You can use /\Ab/.method(:=~).

You can use /\Ab/.method(:=~).

This is more typing than just using '{ |x| x =~ /\Ab/ }'.

I like this idea, it's something I have to do quite often.

I've also wanted a 'grep_v' method before, but reject with Regexp#to_proc would suit perfectly.

[regularfry (Alex Young)]

On 19/02/2013 07:58, rklemme (Robert Klemme) wrote:

Issue #7883 has been reported by rklemme (Robert Klemme).

Lovely. Definite +1 from me (for whatever that's worth).

--
Alex

[rklemme (Robert Klemme)]

nobu (Nobuyoshi Nakada) wrote:

You can use /\Ab/.method(:=~).

Yes, but that's not the point. In that case I'd prefer the real block as Charlie suggested.

[prijutme4ty (Ilya Vorontsov)]

Cool idea. +1

[judofyr (Magnus Holm)]

class Regexp
  def to_proc; lambda {|s| self =~ s} end
end

A more general case:

class Object
  def to_proc
    lambda { |other| self === other }
  end
end

(Not that I think this is going to be accepted…)

[injekt (Lee Jarvis)]

I like this, and I especially like the modified version from judofyr. You could also do ["abc", 3, "foo", etc].find(&Integer) which is neat, but maybe that's a push in the wrong direction, I'm not sure.

[goshakkk (Gosha Arinich)]

Nice one, +1 on it as well.

[jjyr (Jinyang Jiang)]

judofyr (Magnus Holm) wrote:

class Regexp
  def to_proc; lambda {|s| self =~ s} end
end

A more general case:

class Object
  def to_proc
    lambda { |other| self === other }
  end
end

(Not that I think this is going to be accepted…)

cool!

[ko1 (Koichi Sasada)]

• Assignee set to matz (Yukihiro Matsumoto)
• Target version changed from 1.9.3 to 2.6

[rklemme (Robert Klemme)]

Was any of the two suggested forms ever realized?

[nobu (Nobuyoshi Nakada)]

• Description updated (diff)

Nothing, and line 10 is possible with grep

%w{foo bar baz}.grep(/\Ab/) {|s|break s}

Anyway, a patch.

diff --git i/re.c w/re.c
index 1a0f0b1..d345f17 100644
--- i/re.c
+++ w/re.c
@@ -801,6 +801,19 @@ rb_reg_named_captures(VALUE re)
     return hash;
 }
 
+static VALUE
+rb_reg_to_proc(VALUE re)
+{
+    const ID id = 1;
+    VALUE proc = rb_attr_get(re, id);
+    if (NIL_P(proc)) {
+	proc = rb_obj_method(re, ID2SYM(rb_intern("=~")));
+	proc = rb_funcall(proc, rb_intern("to_proc"), 0, 0);
+	rb_ivar_set(re, id, proc);
+    }
+    return proc;
+}
+
 static int
 onig_new_with_source(regex_t** reg, const UChar* pattern, const UChar* pattern_end,
 	  OnigOptionType option, OnigEncoding enc, const OnigSyntaxType* syntax,
@@ -3660,6 +3673,7 @@ Init_Regexp(void)
     rb_define_method(rb_cRegexp, "fixed_encoding?", rb_reg_fixed_encoding_p, 0);
     rb_define_method(rb_cRegexp, "names", rb_reg_names, 0);
     rb_define_method(rb_cRegexp, "named_captures", rb_reg_named_captures, 0);
+    rb_define_method(rb_cRegexp, "to_proc", rb_reg_to_proc, 0);
 
     /* see Regexp.options and Regexp.new */
     rb_define_const(rb_cRegexp, "IGNORECASE", INT2FIX(ONIG_OPTION_IGNORECASE));

[rklemme (Robert Klemme)]

Nobuyoshi Nakada wrote:

Nothing, and line 10 is possible with grep

%w{foo bar baz}.grep(/\Ab/) {|s|break s}

Yes, but a bit lengthy.

Anyway, a patch.

Cool! Thanks a lot!

[sawa (Tsuyoshi Sawada)]

If the proposed patch to issue #9602 gets accepted, then that will cover this issue in a more general way.

[naruse (Yui NARUSE)]

• Target version deleted (2.6)

[hsbt (Hiroshi SHIBATA)]

• Status changed from Open to Assigned