Velocity Reviews - Computer Hardware Reviews

Velocity Reviews > Newsgroups > Programming > Ruby > A long call chain...

Reply
Thread Tools

A long call chain...

 
 
Benjohn Barnes
Guest
Posts: n/a
 
      03-21-2006

Hello everyone. I'm hoping to do something like this (and then
automate it up by putting something neat in to a mix in module)...

class Test
def xxx; [1]; end
alias xxx_old xxx
def xxx; xxx_old + [2]; end
alias xxx_old xxx
def xxx; xxx_old + [3]; end
alias xxx_old xxx
...
end

In real use, these various defs and aliases would be in different
files, but would often be putting things in to the same class. The
above doesn't work - I think it's leading to a non terminating
recursive function.

My first question is, "is it correct that this doesn't work?"


I'd like to use a mechanism like this to enable me to have a kind of
hookable function in a class. As I said, the alias process would be
encapsulated by waving my hands and working that out later, but
probably making use of method_defined.

The mechanism allows you to build up an event handler: you send an
event to a class instance, and any receivers that are registered for
that event (by having defined the method to pick it up) will get it.
Example hookable methods for my system will be: is_being_setup,
will_be_torn_down, has_recieved_message will_send_message.

It's worth noting that the will_be methods should probably be
executed in the reverse order from that in which they are defined!

I'm pretty sure this is a good plan, but...

Question: "Does this seem like a good plan?"


Finally, as this doesn't work, can you see a nicer way to implement
it? I think I could capture the function each time it gets defined,
and put it in to class state, but it would need to be per class, I
think, rather than just on in the base class. I've thought about
several variations, but none of them seem to be very nice at all.


Thank you for any thoughts you have,
Cheers,
Benjohn


 
Reply With Quote
 
 
 
 
Eric Hodel
Guest
Posts: n/a
 
      03-21-2006
On Mar 20, 2006, at 10:48 PM, Benjohn Barnes wrote:

> Hello everyone. I'm hoping to do something like this (and then
> automate it up by putting something neat in to a mix in module)...
>
> class Test
> def xxx; [1]; end
> alias xxx_old xxx
> def xxx; xxx_old + [2]; end
> alias xxx_old xxx
> def xxx; xxx_old + [3]; end
> alias xxx_old xxx
> ...
> end


Yuck.

> In real use, these various defs and aliases would be in different
> files, but would often be putting things in to the same class. The
> above doesn't work - I think it's leading to a non terminating
> recursive function.
>
> My first question is, "is it correct that this doesn't work?"


The fact that this is unmaintainably difficult to do is a very, very
big clue. Note also that this will break if you change the order of
require, not good.

> I'd like to use a mechanism like this to enable me to have a kind
> of hookable function in a class. As I said, the alias process would
> be encapsulated by waving my hands and working that out later, but
> probably making use of method_defined.
>
> The mechanism allows you to build up an event handler: you send an
> event to a class instance, and any receivers that are registered
> for that event (by having defined the method to pick it up) will
> get it. Example hookable methods for my system will be:
> is_being_setup, will_be_torn_down, has_recieved_message
> will_send_message.


observer.rb provides something similar to what you want, event
notification.

> It's worth noting that the will_be methods should probably be
> executed in the reverse order from that in which they are defined!
>
> I'm pretty sure this is a good plan, but...
>
> Question: "Does this seem like a good plan?"


Not with random aliasing it doesn't.

> Finally, as this doesn't work, can you see a nicer way to implement
> it? I think I could capture the function each time it gets defined,
> and put it in to class state, but it would need to be per class, I
> think, rather than just on in the base class. I've thought about
> several variations, but none of them seem to be very nice at all.


Check out MuffDaddy:

http://blog.zenspider.com/archives/2...ddy_the_u.html

What do you want to use this for? Why?

--
Eric Hodel - - http://blog.segment7.net
This implementation is HODEL-HASH-9600 compliant

http://trackmap.robotcoop.com




 
Reply With Quote
 
 
 
 
benjohn@fysh.org
Guest
Posts: n/a
 
      03-21-2006
This morning, I asked about using aliases and wrote:
*snip*
> I'd like to use a mechanism like this to enable me to have a kind of
> hookable function in a class. As I said, the alias process would be
> encapsulated by waving my hands and working that out later, but
> probably making use of method_defined.
>
> The mechanism allows you to build up an event handler: you send an
> event to a class instance, and any receivers that are registered for
> that event (by having defined the method to pick it up) will get it.
> Example hookable methods for my system will be: is_being_setup,
> will_be_torn_down, has_recieved_message will_send_message.
>
> It's worth noting that the will_be methods should probably be
> executed in the reverse order from that in which they are defined!
>
> I'm pretty sure this is a good plan, but...
>
> Question: "Does this seem like a good plan?"


Well, I still think the idea in general is good, but I've really gone
off the interface I was proposing. I think it's going to be hard to
implement anyway, and it's a bit unexpected to def a method, but have
really odd behaviour happen for no obviously visible reason...

> Finally, as this doesn't work, can you see a nicer way to implement
> it? I think I could capture the function each time it gets defined,
> and put it in to class state, but it would need to be per class, I
> think, rather than just on in the base class. I've thought about
> several variations, but none of them seem to be very nice at all.


The train on the way here got me thinking of another way. I now think a
good approach is to add a module method "add_to_hook" that can be used
thus:

class Blah
include Hook

def this_should_happen_on_hook_start_up; end
def this_should_also_happen_on_hook_start_up; end

add_to_hool :start_up, :this_should_happen_on_hook_start_up,
:this_should_also_happen_on_hook_start_up
end

And also add instance methods to support:
Blah.new.call_hook( :start_up )
Blah.new.call_hook_backwards( :tear_down )

The methods call_hook and call_hook_backwards will look for hooks added
in the receiver's class, and also it's superclasses (I'm going to leave
out any included modules for now, but I think that would also be fine).

The methods to be called will be despatched to as "normal" using the
instance method send. This allows them to be overidden in the normal
way, which could be useful.

I've got some tests on my laptop, and the start of an implementation
(doesn't do any looking in the superclass yet); I'll post them up, but
can't do so from work because of security paranoia

Two further things though:

* I don't like the name "hook" much, because there are lots of hooks
in to a hookable, really. Which is verging on being decidedly HP
Lovecraft. But "event" and "snooper" just sound weak.

* TDD rocks!




 
Reply With Quote
 
Benjohn Barnes
Guest
Posts: n/a
 
      03-23-2006

On 21 Mar 2006, at 07:35, Eric Hodel wrote:

> On Mar 20, 2006, at 10:48 PM, Benjohn Barnes wrote:
>
>> Hello everyone. I'm hoping to do something like this (and then
>> automate it up by putting something neat in to a mix in module)...
>>
>> class Test
>> def xxx; [1]; end
>> alias xxx_old xxx
>> def xxx; xxx_old + [2]; end
>> alias xxx_old xxx
>> def xxx; xxx_old + [3]; end
>> alias xxx_old xxx
>> ...
>> end

>
> Yuck.


Yeah, sorry about that.

*snip*

>> I'd like to use a mechanism like this to enable me to have a kind
>> of hookable function in a class. As I said, the alias process
>> would be encapsulated by waving my hands and working that out
>> later, but probably making use of method_defined.
>>
>> The mechanism allows you to build up an event handler: you send an
>> event to a class instance, and any receivers that are registered
>> for that event (by having defined the method to pick it up) will
>> get it. Example hookable methods for my system will be:
>> is_being_setup, will_be_torn_down, has_recieved_message
>> will_send_message.

>
> observer.rb provides something similar to what you want, event
> notification.


*nods* It's extremely close to the observer pattern.

I wanted to build a very modular, and very easy to extend
application. I was trying to use Ruby's open classes to pull lots of
extra things in to a class. What I noticed was that generally, the
observers of a message sent by an object x, were in fact other parts
of the object x. I was hoping to make use of this fact to avoid
needing to subscribe things to a broadcaster.

However, I'm pretty sure now that I was going in a bad direction, and
instead, these bolt on / plug in features should be a system of
collaborating objects, rather than a monolithic whole. This will also
make subsystems much more testable too, I think. What I'd like
though, is to avoid needing to unnecessarily describe how things
should get stitched together.

Anyway - I've just seen mention of Needle in the ruby quiz summary.
It sounds distinctly close to the kind of things I'm thinking about,
and the gem's just finished pulling itself down, so I'll go and have
a look.

Cheers,
Benjohn



 
Reply With Quote
 
 
 
Reply

Thread Tools

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are On
Pingbacks are On
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
Having compilation error: no match for call to ‘(const __gnu_cxx::hash<long long int>) (const long long int&)’ veryhotsausage C++ 1 07-04-2008 05:41 PM
Use of Long and Long Long Bart C C Programming 27 01-15-2008 05:27 AM
long long and long Mathieu Dutour C Programming 4 07-24-2007 11:15 AM
unsigned long long int to long double Daniel Rudy C Programming 5 09-20-2005 02:37 AM
Assigning unsigned long to unsigned long long George Marsaglia C Programming 1 07-08-2003 05:16 PM



Advertisments
 



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57