What is the proper term for things like attr :<id> and belongs_to
:<model>, etc. How do you code them? I tried googling around, but
since I don't know what to call it, its a little tough. I looked at
the ActiveRecord code for a few minutes and it looks like they're just
class methods, but I don't understand how ruby interprets it i guess.
A similar thing has just been discussed on the RubyOnRails list, and
James Britt gave a good example which I'll paste here. It's
implemented (in this case at least) using "meta programming" which is
something I'm just reading up on at the moment, because it foxed me
to!!
Here is James' example:
A simple example
# Meta-programming
# Add a public array property to an object instance
class Meta
def self.has_many( sym )
attr_tmp = "def #{sym.to_s} \n"
attr_tmp << " @#{sym.to_s} = unless @#{sym.to_s}\n"
attr_tmp << " @#{sym.to_s}; end"
eval ( attr_tmp )
end
end
class Foo < Meta
has_many :foo
has_many :baz
end
f = Foo.new
# See what methds the new object has ...
puts (f.methods - Object.methods ).inspect # ["foo", "baz"]
f.foo.push "X"
f.baz << "hello"
p f.foo.first # "X"
p f.baz # ["hello"]
James Britt
···
On Wed, 30 Mar 2005 02:57:58 +0900, Luke Renn <goseigen@comcast.net> wrote:
What is the proper term for things like attr :<id> and belongs_to
:<model>, etc. How do you code them? I tried googling around, but
since I don't know what to call it, its a little tough. I looked at
the ActiveRecord code for a few minutes and it looks like they're just
class methods, but I don't understand how ruby interprets it i guess.
A similar thing has just been discussed on the RubyOnRails list, and
James Britt gave a good example which I'll paste here. It's
implemented (in this case at least) using "meta programming" which is
something I'm just reading up on at the moment, because it foxed me
to!!
Here is James' example:
A simple example
# Meta-programming
# Add a public array property to an object instance
class Meta
def self.has_many( sym )
attr_tmp = "def #{sym.to_s} \n"
attr_tmp << " @#{sym.to_s} = unless @#{sym.to_s}\n"
attr_tmp << " @#{sym.to_s}; end"
eval ( attr_tmp )
end
end
class Foo < Meta
has_many :foo
has_many :baz
end
f = Foo.new
# See what methds the new object has ...
puts (f.methods - Object.methods ).inspect # ["foo", "baz"]
f.foo.push "X"
f.baz << "hello"
p f.foo.first # "X"
p f.baz # ["hello"]
James Britt
···
On Wed, 30 Mar 2005 02:57:58 +0900, Luke Renn <goseigen@comcast.net> wrote:
What is the proper term for things like attr :<id> and belongs_to
:<model>, etc. How do you code them? I tried googling around, but
since I don't know what to call it, its a little tough. I looked at
the ActiveRecord code for a few minutes and it looks like they're just
class methods, but I don't understand how ruby interprets it i guess.
You're absolutely correct that they're just class methods. In Ruby
(unlike many other languages) the code that defines a class is much
more than just a "definition". The code between "class Foo ... end" is
executed just like any other part of your program. Perhaps this simple
example using a global variable will help clear it up:
On Wed, 30 Mar 2005 02:57:58 +0900, Luke Renn <goseigen@comcast.net> wrote:
the ActiveRecord code for a few minutes and it looks like they're just
class methods, but I don't understand how ruby interprets it i guess.
--
Regards,
John Wilger
-----------
Alice came to a fork in the road. "Which road do I take?" she asked.
"Where do you want to go?" responded the Cheshire cat.
"I don't know," Alice answered.
"Then," said the cat, "it doesn't matter."
- Lewis Carrol, Alice in Wonderland
I am not a Ruby expert (yet :-), but the first thing to realize is
that attr, etc are just normal methods. And remembering I am far from
a Ruby expert, I wrote the following code for personal exploration on
the same topic. It adds attr_class_(reader|writer|accessor) methods to
Module so that any other class or module can then provide easy access
to class data variables. Note all it does use class eval to "write"
the access methods -- I do not know if this is what module does or if
there is a substantially better technique. I would prefer to have used
a closure over class_eval, but could not figure out how (if it is
possible).
Patrick
class Module
def attr_class_reader(sym)
self.class_eval <<ATTR
def self.#{sym}
@@#{sym}
end
ATTR
end
def attr_class_writer(sym)
self.class_eval <<ATTR
def self.#{sym}=(new_value)
@@#{sym} = new_value
end
ATTR
end
def attr_class_accessor(sym)
attr_class_reader(sym)
attr_class_writer(sym)
end
end
···
On Wed, 30 Mar 2005 02:57:58 +0900, Luke Renn <goseigen@comcast.net> wrote:
What is the proper term for things like attr :<id> and belongs_to
:<model>, etc. How do you code them? I tried googling around, but
since I don't know what to call it, its a little tough. I looked at
the ActiveRecord code for a few minutes and it looks like they're just
class methods, but I don't understand how ruby interprets it i guess.
Parentheses for method calls are optional for Ruby.
So
attr :foo
is the same as
attr(:foo)
:foo is a Symbol - similar to a String. You can also write
attr('foo')
....since the 'attr' method takes either a Symbol or a String.
If you don't know why Symbol exists (why we don't just use String all the
time), search the archives of this list, and ask again if you're still not
sure.
(The rest of the reason is explained but other people in this thread: Ruby
evaluates statements/expressions "as it sees them".)
···
In article <20050329180034.GA4745@trico.localnet>, Luke Renn wrote:
What is the proper term for things like attr :<id> and belongs_to
:<model>, etc. How do you code them? I tried googling around, but
since I don't know what to call it, its a little tough. I looked at
the ActiveRecord code for a few minutes and it looks like they're just
class methods, but I don't understand how ruby interprets it i guess.
* John Wilger <johnwilger@gmail.com> [2005-03-30 03:09:57 +0900]:
more than just a "definition". The code between "class Foo ... end" is
executed just like any other part of your program. Perhaps this simple
That's what i was thinking. So anything in between class Foo ... end
is like a static { } block in a java class. Thanks to both of you for
clearing that up.
# Meta-programming
# Add a public array property to an object instance
class Meta
def self.has_many( sym )
attr_tmp = "def #{sym.to_s} \n"
attr_tmp << " @#{sym.to_s} = unless @#{sym.to_s}\n"
attr_tmp << " @#{sym.to_s}; end"
eval ( attr_tmp )
end
end
And here it is without an eval:
class Meta
def self.has_many(sym)
ivar = :"@#{sym}"
define_method(sym.to_sym) do
new_value = instance_variable_get(ivar) ||
instance_variable_set(ivar, new_value)
new_value
end
end
end
class Meta
def self.has_many( sym )
attr_tmp = "def #{sym.to_s} \n"
attr_tmp << " @#{sym.to_s} = unless @#{sym.to_s}\n"
attr_tmp << " @#{sym.to_s}; end"
eval ( attr_tmp )
end
end
class Meta
def self.has_many(sym)
ivar = :"@#{sym}"
define_method(sym.to_sym) do
new_value = instance_variable_get(ivar) ||
instance_variable_set(ivar, new_value)
new_value
end
end
end
What is the difference between these two approaches?
* Is one more 'culturally correct'?
* Does one perform better?
* Is one more future-proof?
Florian's will only work on Ruby 1.8 or later and avoids some of the
concerns that people have around "eval".
-austin
···
On Wed, 30 Mar 2005 05:03:48 +0900, B. K. Oxley (binkley) <binkley@alumni.rice.edu> wrote:
Florian Gross wrote:
> class Meta
> def self.has_many(sym)
> ivar = :"@#{sym}"
> define_method(sym.to_sym) do
> new_value = instance_variable_get(ivar) ||
> instance_variable_set(ivar, new_value)
> new_value
> end
> end
> end
What is the difference between these two approaches?
class Meta
def self.has_many( sym )
attr_tmp = "def #{sym.to_s} \n"
attr_tmp << " @#{sym.to_s} = unless @#{sym.to_s}\n"
attr_tmp << " @#{sym.to_s}; end"
eval ( attr_tmp )
end
end
class Meta
def self.has_many(sym)
ivar = :"@#{sym}"
define_method(sym.to_sym) do
new_value = instance_variable_get(ivar) ||
instance_variable_set(ivar, new_value)
new_value
end
end
end
What is the difference between these two approaches?
* Is one more 'culturally correct'?
* Does one perform better?
* Is one more future-proof?
The second approach would generally be considered 'better'. eval must be
used with care (security issues, debugging is harder, etc.). The second
approach doesn't use eval so that's why it would generally be preferred.
Performance differences between the two are probably negligable. I
suspect that either approach would work in the future.
Sometimes you need to use eval (or it's cousins instance_eval and
class_eval - the latter should probably be preferred where possible) and
it's great to have it's power available so I'm not saying you should
_never_ use eval.
Phil
···
B. K. Oxley (binkley) <binkley@alumni.rice.edu> wrote:
What is the difference between these two approaches?
* Is one more 'culturally correct'?
* Does one perform better?
* Is one more future-proof?
James prefers Florian's way for real-life usage.
James wrote his as a sort of "Here's a simple example that may help show what sort stuff is going on" thing.
In general, 'eval' is risky because, if you're not careful, you'll execute the wrong raw code. It can also be harder to debug because you're looking at code + code-in-strings, so the actual intent may be obscured.
The eval example is perhaps more useful for someone looking to hack around to see what else can be done using assorted meta-coding and such, but overall you are probably better off if, once you've decide what to accomplish, you find a non-eval way to do it.
So does there exist (as I cannot find it) a way to access @@vars via
symbols in a similar manner? @vars tied to the class are easy enough
(and can often fill the same niche), but I cannot find a method
similar to instance_variable_XXX for @@ vars.
Thanks
Patrick
···
On Wed, 30 Mar 2005 05:30:04 +0900, Austin Ziegler <halostatue@gmail.com> wrote:
On Wed, 30 Mar 2005 05:03:48 +0900, B. K. Oxley (binkley) > <binkley@alumni.rice.edu> wrote:
> Florian Gross wrote:
> > class Meta
> > def self.has_many(sym)
> > ivar = :"@#{sym}"
> > define_method(sym.to_sym) do
> > new_value = instance_variable_get(ivar) ||
> > instance_variable_set(ivar, new_value)
> > new_value
> > end
> > end
> > end
> What is the difference between these two approaches?
Florian's will only work on Ruby 1.8 or later and avoids some of the
concerns that people have around "eval".
I would take a stab, that minimally "public data" (e.g. web
processing) situation there is a significant code injection risk that
is minimized in the block based code.
Patrick
···
On Wed, 30 Mar 2005 05:32:31 +0900, B. K. Oxley (binkley) <binkley@alumni.rice.edu> wrote:
Florian's will only work on Ruby 1.8 or later and avoids some of the
concerns that people have around "eval".
Being new to Ruby, what sorts of concerns?
Two main issues come to mind:
1) security. What is someone malicious code that gets eval'ed?
"system(\"rm -f *\")"
2) debugging. It can be difficult to debug code that depends on eval.
Errors get reported, but you've got to match line numbers of the string
passed in to the eval function and this can be difficult.
Phil
···
B. K. Oxley (binkley) <binkley@alumni.rice.edu> wrote:
So does there exist (as I cannot find it) a way to access @@vars via
symbols in a similar manner? @vars tied to the class are easy enough
(and can often fill the same niche), but I cannot find a method
similar to instance_variable_XXX for @@ vars.
No, probably because they are not supposed to be used too frequently.
But using just instance_eval("@@" + name) is still better than evaling the whole method definition.
1) security. What is someone malicious code that gets eval'ed? "system(\"rm -f *\")"
2) debugging. It can be difficult to debug code that depends on eval. Errors get reported, but you've got to match line numbers of the string passed in to the eval function and this can be difficult.
3) editing. If your editor understands Ruby syntax, it will highlight and indent the code in a block passed to define_method. But editing code encased in quoted strings makes your editor dumb.