Hi list

the recent thread about all_indicies has somehow made me think that I
need a more general inject.
Here is my first shot on it:
http://pastie.org/597659
But how to call that beast, #inject_with is somehow not really what I like.
Actually I think #inject would be a nice name, but it is already taken.

Cheers
Robert

-- 
If you tell the truth you don't have to remember anything.
--
Samuel Clemens (some call him Mark Twain)

2009/8/28 Robert Dober <robert.dober@gmail.com>:
> the recent thread about all_indicies has somehow made me think that I
> need a more general inject.
> Here is my first shot on it:
> http://pastie.org/597659

I didn't know StopIteration yet.  That's nice.

Your use case is a bit unclear to me.  Also, I do not see why you need
to pass variables via the block.  Somehow that looks suspiciously more
complex to me than it could be.  A few ideas

class String
  def indices rgx, idx = 0
    r = []
    idx = index rgx, idx

    while idx
      r << idx
      idx = index rgx, idx.succ
    end

    r
  end

  def indices2 rgx, idx = 0
    idx = index rgx, idx

    while idx
      yield idx
      idx = index rgx, idx.succ
    end

    self
  end

  def indices3 rgx, idx = 0
    scan rgx do
      i = $`.length
      yield i if i >= idx
    end
  end

  def indices4 rgx, idx = 0
    to_enum(:scan, rgx).inject [] do |r,|
      i = $~.offset(0).first
      yield i if i >= idx
    end
  end
end

irb(main):154:0* "abcabc".indices("a",1)
=> [3]
irb(main):155:0> "abcabc".indices2("a",1) {|i| p i}
"a3
bcabc".indices3("a",1) {|i| p i}
=> "abcabc"
irb(main):156:0> "abcabc".indices3("a",1) {|i| p i}
"abcabc".indices4("a",1) {|i| p i}
3
=> "abcabc"
irb(main):157:0> "abcabc".indices4("a",1) {|i| p i}
3
=> 3
irb(main):158:0>

Ok, I admit, I got carried away. :-)

> But how to call that beast, #inject_with is somehow not really what I like.
> Actually I think #inject would be a nice name, but it is already taken.

Only few come to mind: roll, rotate, circulate

Kind regards

robert

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

On Fri, Aug 28, 2009 at 3:17 PM, Robert
Klemme<shortcutter@googlemail.com> wrote:
<snip>
> Ok, I admit, I got carried away. :-)
Yup, but I guess we all like to read your code. Well I do not care
about indicies, it was just the trigger for: "I want a thing that
passes its results into itself".

Maybe this usecase is more appealing?

http://pastie.org/597712

and of course a logical extension into Enumerable would be

http://pastie.org/597718

R.


-- 
If you tell the truth you don't have to remember anything.
--
Samuel Clemens (some call him Mark Twain)

Hi,

Am Freitag, 28. Aug 2009, 22:17:21 +0900 schrieb Robert Klemme:
> 2009/8/28 Robert Dober <robert.dober@gmail.com>:
> > http://pastie.org/597659
> 
> I didn't know StopIteration yet.  That's nice.

I didn't know it either.  Is there anything that `break' cannot
do better?

Bertram


-- 
Bertram Scharpf
Stuttgart, Deutschland/Germany
http://www.bertram-scharpf.de

On Fri, Aug 28, 2009 at 11:22 AM, Bertram
Scharpf<lists@bertram-scharpf.de> wrote:
> Am Freitag, 28. Aug 2009, 22:17:21 +0900 schrieb Robert Klemme:
>> 2009/8/28 Robert Dober <robert.dober@gmail.com>:
>> > http://pastie.org/597659
>>
>> I didn't know StopIteration yet. =A0That's nice.
>
> I didn't know it either.

Neither did I.  It's a Ruby 1.9 thing, backported to 1.8.7 some of the
other stuff in that snippet like the block arg syntax is Ruby 1.9
only.

--=20
Rick DeNatale

Blog: http://talklikeaduck.denhaven2.com/
Twitter: http://twitter.com/RickDeNatale
WWR: http://www.workingwithrails.com/person/9021-rick-denatale
LinkedIn: http://www.linkedin.com/in/rickdenatale

Hi,

I have a string literal in this form:

  taxonomy =3D %q{
    +AttributeControl
      NMAttributeControl.*
    +NetworkControl
      NMNetworkControl*.*
    +TimelineControl
      NMTimelineControl*.*
    +CurveControl
      NMCurveControl*.*
    *.*
  }


Is it possible to somehow specify it such that there are (automatically) ne=
wline characters at the end of each line?

Cheers,
James

On 28.08.2009 17:22, Bertram Scharpf wrote:
  > Am Freitag, 28. Aug 2009, 22:17:21 +0900 schrieb Robert Klemme:
>> 2009/8/28 Robert Dober <robert.dober@gmail.com>:
>>> http://pastie.org/597659
>> I didn't know StopIteration yet.  That's nice.
> 
> I didn't know it either.  Is there anything that `break' cannot
> do better?

Multilevel exit like you can also do with catch throw (see Find.find).

Kind regards

	robert

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

James French schrieb:
> Hi,
>
> I have a string literal in this form:
>
>   taxonomy = %q{
>     +AttributeControl
>       NMAttributeControl.*
>     +NetworkControl
>       NMNetworkControl*.*
>     +TimelineControl
>       NMTimelineControl*.*
>     +CurveControl
>       NMCurveControl*.*
>     *.*
>   }
>
>
> Is it possible to somehow specify it such that there are (automatically) newline characters at the end of each line?
>
> Cheers,
> James
>
>   
Hi,

there ARE already Newlines in your string:

  puts taxonomy.scan(/\n/).size #=> 10

What's your problem?

Marvin

Boy do I feel stupid. taxonomy.include?('\n') was returning false. Wrong qu=
oting... Cheers for the nudge.


> -----Original Message-----
> From: Quintus [mailto:sutniuq@gmx.net]
> Sent: 28 August 2009 18:01
> To: ruby-talk ML
> Subject: Re: new lines in string literals
>=20
> James French schrieb:
> > Hi,
> >
> > I have a string literal in this form:
> >
> >   taxonomy =3D %q{
> >     +AttributeControl
> >       NMAttributeControl.*
> >     +NetworkControl
> >       NMNetworkControl*.*
> >     +TimelineControl
> >       NMTimelineControl*.*
> >     +CurveControl
> >       NMCurveControl*.*
> >     *.*
> >   }
> >
> >
> > Is it possible to somehow specify it such that there are
> (automatically) newline characters at the end of each line?
> >
> > Cheers,
> > James
> >
> >
> Hi,
>=20
> there ARE already Newlines in your string:
>=20
>   puts taxonomy.scan(/\n/).size #=3D> 10
>=20
> What's your problem?
>=20
> Marvin

On Fri, Aug 28, 2009 at 6:55 PM, Robert
Klemme<shortcutter@googlemail.com> wrote:
> On 28.08.2009 17:22, Bertram Scharpf wrote:
> =A0> Am Freitag, 28. Aug 2009, 22:17:21 +0900 schrieb Robert Klemme:
>>>
>>> 2009/8/28 Robert Dober <robert.dober@gmail.com>:
>>>>
>>>> http://pastie.org/597659
>>>
>>> I didn't know StopIteration yet. =A0That's nice.
>>
>> I didn't know it either. =A0Is there anything that `break' cannot
>> do better?
>
> Multilevel exit like you can also do with catch throw (see Find.find).
Yep, by using "raise StopIteration" I make my intent a little bit
clearer and it is less fragile
in case I am getting embedded into other blox. But that is not the
case here, I agree.

I think it might also having been built for Enumerator#next

loop do
    ...
    an_enum.next # might raise StopIteration
    ...
end

HTH
R.

Hi,

Am Samstag, 29. Aug 2009, 01:55:26 +0900 schrieb Robert Klemme:
> On 28.08.2009 17:22, Bertram Scharpf wrote:
>  > Am Freitag, 28. Aug 2009, 22:17:21 +0900 schrieb Robert Klemme:
>>> 2009/8/28 Robert Dober <robert.dober@gmail.com>:
>>>> http://pastie.org/597659
>>> 
>>> I didn't know StopIteration yet.  That's nice.
>> 
>> I didn't know it either.  Is there anything that `break' cannot
>> do better?
>
> Multilevel exit

No, it exits just one level:

  loop { puts "x" ; loop { puts "y" ; raise StopIteration } }

Bertram


-- 
Bertram Scharpf
Stuttgart, Deutschland/Germany
http://www.bertram-scharpf.de

On Fri, Aug 28, 2009 at 7:43 PM, Bertram
Scharpf<lists@bertram-scharpf.de> wrote:
>
> =A0loop { puts "x" ; loop { puts "y" ; raise StopIteration } }
Wait a second, break exits a block,
raise gets to the next,exception handler, in your case the one in the
inner loop; thus you need to do the following

loop  do
  puts "x"
  lambda{ puts "y"; raise StopIteration}[]
end

to see what is going on

In other words you did

begin
   while true
      puts "x"
      begin
        while true
           puts "y"
           raise SI
        end
     rescue SI
     end
  end
rescue SI
end


R.

Sorry for re-phrasing my question, but it just came back to me, from
where I had this idea:
Clojure's iterate

Do you think iterate is a good name for this?

Cheers
Robert

On 28.08.2009 19:43, Bertram Scharpf wrote:
> Hi,
> 
> Am Samstag, 29. Aug 2009, 01:55:26 +0900 schrieb Robert Klemme:
>> On 28.08.2009 17:22, Bertram Scharpf wrote:
>>  > Am Freitag, 28. Aug 2009, 22:17:21 +0900 schrieb Robert Klemme:
>>>> 2009/8/28 Robert Dober <robert.dober@gmail.com>:
>>>>> http://pastie.org/597659
>>>> I didn't know StopIteration yet.  That's nice.
>>> I didn't know it either.  Is there anything that `break' cannot
>>> do better?
>> Multilevel exit
> 
> No, it exits just one level:
> 
>   loop { puts "x" ; loop { puts "y" ; raise StopIteration } }

Well, we're both right: it won't exit multiple *loop* levels but it will 
exit multiple *stack frames*, i.e. method and block calls (not 
containing loops of course).  That's what I had in mind.

Your point demonstrates though that catch throw might be superior 
because you explicitly declare the point of return.

Kind regards

	robert

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

On Sat, Aug 29, 2009 at 11:20 AM, Robert
Klemme<shortcutter@googlemail.com> wrote:
> On 28.08.2009 19:43, Bertram Scharpf wrote:

> Your point demonstrates though that catch throw might be superior because
> you explicitly declare the point of return.
I just think they are two different tools for two different tasks.

   raise SI:  end an iteration / loop
   throw/catch raise * : Whatever seems appropriate, non local exit
being a possibility

It depends heavily on the *.

I therefore reconsider my confirmation that SI is for non local exit
in general, Betram's examples have made that quite clear.

Cheers
R.