Is it possible to overload constants in Perl 6? Here is the Perl 5 example I'm looking at.
In particular I would like to get a string of the literal value used, e.g. if the code was
my $x = .1e-003 ;
I need ".1e-003" instead of 0.0001.
I just added such a module:
https://github.com/FROGGS/p6-overload-constant
USAGE:
use v6;
sub decimal { $^a.flip }
use overload::constant &decimal;
say .1e-003 # "300-e1."
You can change how a value stringifies by mixing in an appropriate role with the but operator, ie
0.0001 but role { method Str { ".1e-003" } }
which can be shortened to
0.0001 but ".1e-003"
Note that providing a method Stringy instead of Str might actually be more appropriate from a semantic point of view, but I do not think Rakudo as of today handles that distinction correctly in all case.
I don't think anything like overload::constant exists in the spec or the existing libraries, but it should be possible to write it using macros by looking in the AST for the type of literal that you're interested in, and replacing it with an object constructor or whatever you need.
Cleaner, perhaps, would be to wrap each one of these constants in a macro invocation, instead of spanning the whole program with one.
Related
Perl is a bit too forgiving: If you pass extra arguments to subs they are simply ignored.
To avoid this I would like to use prototypes to make sure each sub is given the correct amount of arguments.
This works OK as long as I declare the prototype before using it:
sub mysub($);
sub mysub2($);
mysub(8);
mysub(8,2); # Complain here
sub mysub($) {
mysub2($#);
}
sub mysub2($) {
if($_[0] == 1) {
mysub(2);
}
print $#;
}
But I really hate splitting this up. I would much rather that Perl read the full file to see if there are declarations further down. So I would like to write something like:
use prototypes_further_down; # This does not work
mysub(8);
mysub(8,2); # Complain here
sub mysub($) {
mysub2($#);
}
sub mysub2($) {
if($_[0] == 1) {
mysub(2);
}
print $#;
}
Can I somehow ask Perl to do that?
To avoid this I would like to use prototypes to make sure each sub is given the correct amount of arguments.
No, you would not. Despite the similarity in name, Perl prototypes are not your father's function prototypes. Quoting The Problem with Prototypes (emphasis mine),
Perl 5's prototypes serve two purposes. First, they're hints to the parser to change the way it parses subroutines and their arguments. Second, they change the way Perl 5 handles arguments to those subroutines when it executes them. A common novice mistake is to assume that they serve the same language purpose as subroutine signatures in other languages. This is not true.
In addition to them not having the same intended purpose, bypassing prototypes is trivial, so they provide no actual protection against someone who deliberately wishes to call your code in (what you believe to be) the "wrong" way. As perldoc perlsub tells us,
The function declaration must be visible at compile time. The prototype affects only interpretation of new-style calls to the function, where new-style is defined as not using the & character. In other words, if you call it like a built-in function, then it behaves like a built-in function. If you call it like an old-fashioned subroutine, then it behaves like an old-fashioned subroutine. It naturally falls out from this rule that prototypes have no influence on subroutine references like \&foo or on indirect subroutine calls like &{$subref} or $subref->().
Method calls are not influenced by prototypes either, because the function to be called is indeterminate at compile time, since the exact code called depends on inheritance.
Even if you could get it to complain about mysub(8,2), &mysub(8,2) or $subref = \&mysub; $subref->(8,2) or (if mysub were an object method inside package MyModule) $o = MyModule->new; $o->mysub(8,2) would work without complaint.
If you want to validate how your subs are called using core Perl (prior to 5.20), then you need to perform the validation yourself within the body of the sub. Perl 5.20 and newer have a ("experimental" at the time of this writing) Signatures extension to sub declarations which may work for your purposes, but I've never used it myself, so I can't speak to its effectiveness or limitations. There are also many CPAN modules available for handling this sort of thing, which you can find by doing searches for things like "signature" or "prototype".
Regardless of your chosen approach, you will not be able to get compile-time errors about incorrect function signatures unless you define those signatures before they are used. In cases such as your example, where two subs mutually call each other, this can be accomplished by using a forward declaration to establish its signature in advance:
sub mysub($foo); # Forward declaration
sub mysub2 { mysub(8) }
sub mysub { mysub2('infinite loops ftw!') } # Complete version of the code
Sorry to bother the community for this but I have unfortunately to code in Perl :'(. It is about an OO perl code I want to understand but I am failing to put all the pieces together.
The following is a template of code that represents somehow what I am currently looking at. The following is the class MyClass:
package Namespace::MyClass;
sub new($)
{
my ($class) = #_;
$self = { };
bless ($self, $class);
}
sub init($$)
{
my ($self, $param1) = #_;
$self->{whatever} = ($param1, $param1, $param1);
}
and then the following is a script.pl that supposedly uses the class:
#!/path/to/your/perl
require Namespace::MyClass;
my myClass = new Namespace::MyClass()
myClass->init("data_for_param1");
There may be error but I am interested more in having the following questions answered than having my possibly wrong code corrected:
Questions group 1 : "$" in a sub definition means I need to supply one parameter, right? If so, why does new ask for one and I do not supply it? Has this to do with the call in the script using () or something similar to how Python works (self is implied)?
Question group 2 : is for the same previous reason that the init subroutine (here a method) declares to expect two parameters? If so, is the blessing in some way implying a self is ever passed for all the function in the module?
I ask this because I saw that in non blessed modules one $ = one parameter.
Thank you for your time.
QG1:
Prototypes (like "$") mean exactly nothing in Method calls.
Method calls are not influenced by prototypes either, because the function to be called is indeterminate at compile time, since the exact code called depends on inheritance.
Most experienced Perl folk avoid prototypes entirely unless they are trying to imitate a built-in function. Some PHBs inexperienced in Perl mandate their use under the mistaken idea that they work like prototypes in other languages.
The 1st parameter of a Method call is the Object (Blessed Ref) or Class Name (String) that called the Method. In the case of your new Method that would be 'Namespace::MyClass'.
Word to the wise: Also avoid indirect Method calls. Rewrite your line using the direct Method call as follows: my $myClass = Namespace::MyClass->new;
QG2:
Your init method is getting $myClass as it's 1st parameter because it is what 'called' the method. The 2nd parameter is from the parameter list. Blessing binds the name of the Class to the Reference, so that when a method call is seen, It knows which class in which to start the search for the correct sub. If the correct sub is not immediately found, the search continues in the classes named in the class's #ISA array.
Don't use prototypes! They don't do what you think they do.
Prototypes in Perl are mainly used to allow functions to be defined without the use of parentheses or to allow for functions that take array references to use the array name like pop or push do. Otherwise, prototypes can cause more trouble and heartbreak than experienced by most soap opera characters.
is what you actually want to do validate parameters? if so then that is not the purpose of prototypes. you could try using signatures, but for some reason they are new and still experimental. some consider lack of a stable signatures feature to be a flaw of perl. the alternatives are CPAN and writing code in your subs/methods that explicitly validate the params.
In Javascript, the "for (attr in this)" is often dangerous to use... I agree. That's one reason I like Coffeescript. However, I'm programming in Coffeescript and have a case where I need Javascript's "for (attr in this)". Is there a good way to do this in Coffeescript?
What I am doing now is writing a bunch of logic in embedded raw Javascript, such as:
...coffeescript here...
for (attr in this) {
if (stuff here) {
etc
}
}
It'd be nice to use as little Javascript as possible... any suggestions for how I can achieve this and maximize my use of Coffeescript?
Instead of for item in items which iterates through arrays, you can use for attr, value of object, which works more like for in from JS.
for own attr, value of this
if attr == 'foo' && value == 'bar'
console.log 'Found a foobar!'
Compiled: https://gist.github.com/62860f0c07d60320151c
It accepts both the key and the value in the loop, which is very handy. And you can insert the own keyword right after the for in order to enforce an if object.hasOwnProperty(attr) check which should filter out anything from the prototype that you don't want in there.
Squeegy's answer is correct. Let me just amend it by adding that the usual solution to JavaScript's for...in being "dangerous" (by including prototype properties) is to add a hasOwnProperty check. CoffeeScript can do this automatically using the special own keyword:
for own attr of this
...
is equivalent to the JavaScript
for (attr in this) {
if (!Object.prototype.hasOwnProperty(this, attr)) continue;
...
}
When in doubt about whether you should use for...of or for own...of, it's generally safer to use own.
You can use for x in y or for x of y depending on how you want to interpret a list of elements. The newest version of CoffeeScript aims to solve this problem, and you can read about its new use with an issue (that has since been implemented and closed) here on GitHub
Example: I want to do this:
METHODNAME(5) {
// do something
}
which results in:
- (void)animationStep5 {
// do something
}
Is there any way to do this? Basically, what I need is a way to generate a real source code string before the program is compiled, so the compiler does see - (void)animationStep5...
Or maybe there's something different than a macro, which can help here to auto-generate method names (not at run-time)?
As was already answered here, the objective-C preprocessor is very close to the C one.
You should have a look at the examples posted there, and have a look at C proprocessor. You will simply have to use the ## syntax of the preprocessor, to concatenate the method name, and the number you want.
You can use the concatenation operator
#define METHODNAME(i) -(void)animationStep##i
you can call it like
METHODNAME(5){}
This expands to
-(void)animationStep5{}
Assuming the objective-c preprocessor behaves the same as the standard C one, you can use something like:
#define PASTE(a, b) a##b
#define METHODNAME(n) PASTE(animationStep,n)
to join the required bits together. This means that
METHODNAME(5)
gets translated to
animationStep5
(you may need to add the "void" from your question to the macro definitino depending on exactly what it is you need to do).
Many years ago I remember a fellow programmer counselling this:
new Some::Class; # bad! (but why?)
Some::Class->new(); # good!
Sadly now I cannot remember the/his reason why. :( Both forms will work correctly even if the constructor does not actually exist in the Some::Class module but instead is inherited from a parent somewhere.
Neither of these forms are the same as Some::Class::new(), which will not pass the name of the class as the first parameter to the constructor -- so this form is always incorrect.
Even if the two forms are equivalent, I find Some::Class->new() to be much more clear, as it follows the standard convention for calling a method on a module, and in perl, the 'new' method is not special - a constructor could be called anything, and new() could do anything (although of course we generally expect it to be a constructor).
Using new Some::Class is called "indirect" method invocation, and it's bad because it introduces some ambiguity into the syntax.
One reason it can fail is if you have an array or hash of objects. You might expect
dosomethingwith $hashref->{obj}
to be equal to
$hashref->{obj}->dosomethingwith();
but it actually parses as:
$hashref->dosomethingwith->{obj}
which probably isn't what you wanted.
Another problem is if there happens to be a function in your package with the same name as a method you're trying to call. For example, what if some module that you use'd exported a function called dosomethingwith? In that case, dosomethingwith $object is ambiguous, and can result in puzzling bugs.
Using the -> syntax exclusively eliminates these problems, because the method and what you want the method to operate upon are always clear to the compiler.
See Indirect Object Syntax in the perlobj documentation for an explanation of its pitfalls. freido's answer covers one of them (although I tend to avoid that with explicit parens around my function calls).
Larry once joked that it was there to make the C++ feel happy about new, and although people will tell you not to ever use it, you're probably doing it all the time. Consider this:
print FH "Some message";
Have you ever wondered my there was no comma after the filehandle? And there's no comma after the class name in the indirect object notation? That's what's going on here. You could rewrite that as a method call on print:
FH->print( "Some message" );
You may have experienced some weirdness in print if you do it wrong. Putting a comma after the explicit file handle turns it into an argument:
print FH, "some message"; # GLOB(0xDEADBEEF)some message
Sadly, we have this goofiness in Perl. Not everything that got into the syntax was the best idea, but that's what happens when you pull from so many sources for inspiration. Some of the ideas have to be the bad ones.
The indirect object syntax is frowned upon, for good reasons, but that's got nothing to do with constructors. You're almost never going to have a new() function in the calling package. Rather, you should use Package->new() for two other (better?) reasons:
As you said, all other class methods take the form Package->method(), so consistency is a Good Thing
If you're supplying arguments to the constructor, or you're taking the result of the constructor and immediately calling methods on it (if e.g. you don't care about keeping the object around), it's simpler to say e.g.
$foo = Foo->new(type => 'bar', style => 'baz');
Bar->new->do_stuff;
than
$foo = new Foo(type => 'bar', style => 'baz');
(new Bar)->do_stuff;
Another problem is that new Some::Class happens at run time. If there is an error and you testing never branches to this statement, you never know it until it happens in production. It is better to use Some::Class->new unless you are doing dynamic programing.