[#403837] Why none of the block giving the expected output with the "enumerator"? — Arup Rakshit <lists@...>

Why none of the block giving the expected output with the "enumerator"?

9 messages 2013/02/02

[#403870] Confusion with Enum#with_object block argument construct — Arup Rakshit <lists@...>

C:\>irb

9 messages 2013/02/03

[#403920] Character classes use in Ruby — Love U Ruby <lists@...>

Can anyone help me by giving an explanatory example of each of the

13 messages 2013/02/04

[#403935] How to stop page loading using selenium-web driver? — Love U Ruby <lists@...>

How to stop page loading using selenium-web driver?

11 messages 2013/02/04

[#403972] Ruby could recognize the values when putting into a webpage text filed. — Love U Ruby <lists@...>

SGksCgpJIGhhdmUgc29tZSB2YWx1ZXMgaW4gbXkgRXhjZWwgYXMgYmVsb3c6

9 messages 2013/02/05

[#403986] old syntax? what's going on here — tamouse mailing lists <tamouse.lists@...>

I have this in a _spec.rb file: (a gem I inherited at work)

11 messages 2013/02/05

[#404005] Implementing DRY with a function call — Rob Marshall <lists@...>

Hi,

12 messages 2013/02/05

[#404006] using an instance variable inside a method — FirstName Surname <lists@...>

Hello.

19 messages 2013/02/05

[#404021] Not able get the label text incurred with <input> element — Love U Ruby <lists@...>

I do have a below `HTML`:

18 messages 2013/02/05

[#404025] Symbol.defined? — Student Jr <lists@...>

Symbol deserialization from external sources is now known to be

15 messages 2013/02/06

[#404058] Ruby 1.9.3-p362 on Mac OSX — Peter Bailey <lists@...>

Hi,

13 messages 2013/02/06

[#404082] Problem building Ruby 1.9.3 patchlevel 385 under AIX 7.1 — Ruby Student <ruby.student@...>

Hello World!

10 messages 2013/02/06

[#404101] Issues from an extreme beginner — Bruce Palmer <lists@...>

Hey guys, great to be part of such a great community! I look forward to

19 messages 2013/02/07
[#404104] Re: Issues from an extreme beginner — Bruce Palmer <lists@...> 2013/02/07

Ahh, thank you Matthew! That was just the push I needed!

[#404208] elegant way to determine if something is defined — tamouse mailing lists <tamouse.lists@...>

Something like:

15 messages 2013/02/10

[#404218] Ruby Equivalent to VB's "With"? — Joel Pearson <lists@...>

I've looked around but I couldn't find anything helpful on this,

11 messages 2013/02/10

[#404235] The "ruby way" to do desktop applications? — "guirec c." <lists@...>

Hello,

17 messages 2013/02/11

[#404238] Best books for "advanced" programmers — "guirec c." <lists@...>

Hello,

18 messages 2013/02/11

[#404245] Issue with Excel column values read. — Love U Ruby <lists@...>

Hi,

22 messages 2013/02/11

[#404344] Ruby command line options s and S — Love U Ruby <lists@...>

Can anyone help me to understand the difference between s and S with

15 messages 2013/02/13

[#404386] Re: Ruby command line options s and S — "D. Deryl Downey" <me@...>

Dude!

20 messages 2013/02/14

[#404387] Ruby Multithreaded producer-consumer problem — Abhijit Sarkar <lists@...>

Hi,

26 messages 2013/02/14
[#404896] Re: Ruby Multithreaded producer-consumer problem — Abhijit Sarkar <lists@...> 2013/02/24

Bump!

[#404456] skip iteration in each loop — Saurav Chakraborty <lists@...>

I want to skip iteration for few values depending on dynamic condition.

11 messages 2013/02/15

[#404491] so, what's the proper way to replace funcionality of GOTO ? — "Stu P. D'naim" <lists@...>

I need to make few scripts for tasks I do often manually, but last time

27 messages 2013/02/15
[#404492] Re: [from BASIC to Ruby] so, what's the proper way to replace funcionality of GOTO ? — Love U Ruby <lists@...> 2013/02/15

Stu P. D'naim wrote in post #1097111:

[#404494] Re: [from BASIC to Ruby] so, what's the proper way to replace funcionality of GOTO ? — Ryan Victory <ryan@...> 2013/02/15

Love U Ruby: I'm really not sure what you meant by that response, but

[#404570] What is Ruby's default constructor? — Love U Ruby <lists@...>

Hi,

12 messages 2013/02/17

[#404632] Re: splat operator and Ruby instance variable assignments — Marc Heiler <lists@...>

Ok, understood what the guy wants ...

17 messages 2013/02/19
[#404640] Re: splat operator and Ruby instance variable assignments — Love U Ruby <lists@...> 2013/02/19

Marc Heiler wrote in post #1097736:

[#404645] Re: splat operator and Ruby instance variable assignments — Ryan Davis <ryand-ruby@...> 2013/02/19

[#404646] Re: splat operator and Ruby instance variable assignments — Love U Ruby <lists@...> 2013/02/19

Ryan Davis wrote in post #1097840:

[#404647] Re: splat operator and Ruby instance variable assignments — Peter Hickman <peterhickman386@...> 2013/02/19

On 19 February 2013 20:35, Love U Ruby <lists@ruby-forum.com> wrote:

[#404648] Re: splat operator and Ruby instance variable assignments — Love U Ruby <lists@...> 2013/02/19

Peter Hickman wrote in post #1097848:

[#404696] THE CLASS/OBJECT CHICKEN-AND-EGG PARADOX — "Xavier R." <lists@...>

Hi,

25 messages 2013/02/20
[#404699] Re: THE CLASS/OBJECT CHICKEN-AND-EGG PARADOX — Matt Mongeau <halogenandtoast@...> 2013/02/20

Maybe you could provide more detail about what you are confused about. To

[#404700] Re: THE CLASS/OBJECT CHICKEN-AND-EGG PARADOX — "Xavier R." <lists@...> 2013/02/20

Matt Mongeau wrote in post #1098058:

[#404705] Re: THE CLASS/OBJECT CHICKEN-AND-EGG PARADOX — Matt Mongeau <halogenandtoast@...> 2013/02/20

It's not really a paradox. Take for example

[#404738] backslash substitution — Mario Ruiz <lists@...>

don't know why... but this is not working

18 messages 2013/02/21

[#404809] Difference of 2 dates interms of years. — "Xavier R." <lists@...>

how can we get the experience years between two dates(ex:2012-01-11 to

13 messages 2013/02/22

[#404817] Not able to understand the difference between "||=" and "|=". — "Xavier R." <lists@...>

>> a = []

12 messages 2013/02/22

[#404839] range is not assigning to the splat variable. — Love U Ruby <lists@...>

Why splat variable couldn't take in the below two code the "range" ->

10 messages 2013/02/23

[#404842] Why class returning its own name when "include" statement? — Love U Ruby <lists@...>

I was actually playing around with the class definition return values.

16 messages 2013/02/23
[#404844] Re: Why class returning its own name when "include" statement? — Love U Ruby <lists@...> 2013/02/23

@Stefano Yes you are right. The below code is proved that.

[#404867] how to see the class creation time in Ruby? — Love U Ruby <lists@...>

In Ruby any chance to see the last-modified time of a specific class?

15 messages 2013/02/23

[#404901] Confusion with `nil` value being produced by IRB in case of Array#size manipulation. — Love U Ruby <lists@...>

enum[int] =3D obj =E2=86=92 obj

10 messages 2013/02/24

[#404921] How should I print only the last combination when using Array#combination(n) ? — Love U Ruby <lists@...>

>> a = [1,2,3]

15 messages 2013/02/24

[#405026] Please, help (GCD) greatest common divisor. — Caddy Tonks Lupin <lists@...>

Write a program to read two integers and show their greatest common

17 messages 2013/02/26

[#405059] Does this specific sound library exist? — Dirk Vogel <lists@...>

Hi there,

16 messages 2013/02/26

[#405067] Mac OS 10.8.2 and openssl — "Dr. Hegewald" <hegewald@...>

Hi everybody,

24 messages 2013/02/27

[#405079] Why `10` not returned without the `return` from the block ? — Love U Ruby <lists@...>

CODE - I

10 messages 2013/02/27

[#405107] Object track llist for a particular class. — "Xavier R." <lists@...>

Say I have created more than one instances from a particular class as

13 messages 2013/02/27

[#405145] Discussion on Ruby's `alias` — Tukai Patra <lists@...>

>> class Foo

27 messages 2013/02/28

[#405175] telnet - how to loop through commands listed in a file — Bob Ford <lists@...>

Let me first explain what I'm trying to do. I have written a very

18 messages 2013/02/28

Re: Questionable regex performance when using lazy matching and inverted character classes

From: Robert Klemme <shortcutter@...>
Date: 2013-02-22 14:32:01 UTC
List: ruby-talk #404797
On Fri, Feb 22, 2013 at 1:21 AM, Tikhon B. <lists@ruby-forum.com> wrote:
> Robert Klemme wrote in post #1098199:
>> On Thu, Feb 21, 2013 at 12:12 PM, Tikhon B. <lists@ruby-forum.com>
>> wrote:

>> You are doing the measurement wrong.  You do not only measure matching
>> but also creation of the String.  Given that fact, all your results
>> are moot.
>
> I actively considered this matter when submitting the post, but in the
> end I included the string creating in the test to allow anyone to try
> out the example by copying one line.

Having a second line which defines a variable or constant isn't really
that more inconvenient.

> The time necessary for the actual
> creation process is negligible; I need to run the string creation ten
> thousand times before the result is substantial enough to significantly
> affect the numbers I'm getting.

Still you are measuring one thing but reasoning about another.

> I am not sure what sort of proper and complete tests you are looking
> for.

Tests which measure exactly the functionality you want to scrutinize
not something else.

> Regarding the actual slowdown, the issue comes down to the engine
> repeating the match from the start for every single "a" in the string,
> then failing and trying again with the next letter.

Backtracking.

> I do not see how a
> greedy modifier can help in this situation, and I do not know enough
> about the ruby regex engine to comment on the underlying design.

Mine was a general statement about issues that can be avoided with
greediness modifiers.  I didn't say it's the solution to all
performance issues of this kind.

I had a closer look at your expressions.  They are of course
artificial to demonstrate the effects of backtracking.  I would assume
that one would typically write different expressions in real
applications, for example, using anchoring which dramatically improves
the situation here.

$ ruby rx-bm.rb
                                user     system      total        real
 0 /(?>a[^b]*b)/           11.732000   0.000000  11.732000 ( 11.737671)
 1 /(?>a[^b]*b)/            0.015000   0.000000   0.015000 (  0.003000)
 0 /\A(?>a[^b]*b)/          0.000000   0.000000   0.000000 (  0.002000)
 1 /\A(?>a[^b]*b)/          0.000000   0.000000   0.000000 (  0.003000)
 0 /(?<!a)(?>a[^b]*b)/      0.016000   0.000000   0.016000 (  0.012001)
 1 /(?<!a)(?>a[^b]*b)/      0.000000   0.000000   0.000000 (  0.003000)
 0 /aa*b/                  11.700000   0.000000  11.700000 ( 11.703670)
 1 /aa*b/                   0.000000   0.000000   0.000000 (  0.002000)
 0 /\Aaa*b/                 0.000000   0.000000   0.000000 (  0.002000)
 1 /\Aaa*b/                 0.015000   0.000000   0.015000 (  0.002000)
 0 /(?<!a)aa*b/             0.000000   0.000000   0.000000 (  0.013001)
 1 /(?<!a)aa*b/             0.016000   0.000000   0.016000 (  0.002000)
 0 /a+b/                   11.248000   0.000000  11.248000 ( 11.252644)
 1 /a+b/                    0.000000   0.000000   0.000000 (  0.002000)
 0 /\Aa+b/                  0.000000   0.000000   0.000000 (  0.003000)
 1 /\Aa+b/                  0.000000   0.000000   0.000000 (  0.002000)
 0 /(?<!a)a+b/              0.015000   0.000000   0.015000 (  0.012001)
 1 /(?<!a)a+b/              0.000000   0.000000   0.000000 (  0.003000)
 0 /a.*b/                   4.649000   0.000000   4.649000 (  4.641265)
 1 /a.*b/                   0.000000   0.000000   0.000000 (  0.001000)
 0 /\Aa.*b/                 0.000000   0.000000   0.000000 (  0.001000)
 1 /\Aa.*b/                 0.000000   0.000000   0.000000 (  0.001000)
 0 /(?<!a)a.*b/             0.016000   0.000000   0.016000 (  0.011001)
 1 /(?<!a)a.*b/             0.000000   0.000000   0.000000 (  0.000000)
$ cat -n rx-bm.rb
     1
     2  require 'benchmark'
     3
     4  strings = [
     5    ("a" * 10_000).freeze,
     6    ("a" * 10_000 + "b").freeze
     7  ]
     8
     9  data = [
    10          /(?>a[^b]*b)/,
    11          /\A(?>a[^b]*b)/,
    12          /(?<!a)(?>a[^b]*b)/,
    13
    14          /aa*b/,
    15          /\Aaa*b/,
    16          /(?<!a)aa*b/,
    17
    18          /a+b/,
    19          /\Aa+b/,
    20          /(?<!a)a+b/,
    21
    22          /a.*b/,
    23          /\Aa.*b/,
    24          /(?<!a)a.*b/,
    25  ]
    26
    27  REP = 20
    28
    29  Benchmark.bm 25 do |x|
    30    data.each do |rx|
    31      strings.each_with_index do |s, i|
    32        x.report sprintf("%2d %p", i, rx) do
    33          REP.times do
    34            rx.match s
    35          end
    36        end
    37      end
    38    end
    39  end
$

So, yes, there is a cost for backtracking involved, but it only shows
only in particular situations:
- large inputs
- specific expressions (absence of anchoring, using unlimited
repetition operators vs. limited (e.g. /a{0,10}/).

The phenomenon is well known with regular expression engines I
believe.  The cases you mention seem to be rather the exception than
the common state of affairs.  Which doesn't mean that Ruby's engine
cannot be improved.  I just don't think it's not that dramatic an
issue.

Kind regards

robert

-- 
remember.guy do |as, often| as.you_can - without end
http://blog.rubybestpractices.com/

In This Thread