In ActionScript 3, you can check if a value exists like this:
if (object) {
trace("This object is not null, undefined, or empty!")
}
I frequently use this as a shorthand for if (object != null)
Is there a proper term for evaluating objects for null in this fashion? I suppose it's a matter of the Boolean typecasting rules for the language but I'm not sure if there's a name for the resulting syntax.
If your object is always an actual object reference, e.g. a variable of any object type, then checking if (object) is a valid way to test for nulls. If it's a property of variant type, or a dynamic property that can potentially contain simple values (ints, strings etc) then the proper way to test for null will be explicit conversion, probably even with a strict type check if (object !== null).
Like most computer languages, Ecma-script supports Boolean data types;
values which can be set to true or false. In addition, everything in
JavaScript has an inherent Boolean value, generally known as either
truthy or falsy
list of falsy values (non truthy)
false
0 (zero)
"" (empty string)
null
undefined
NaN (a special Number value meaning Not-a-Number!)
regularly for checking if a variable is null
you may use : (a == null) this statement returns true if a is null, But also return true for all of the above list, because they'r all falsy values and (a == undefined) returns true, even a is not undefined but null or 0 or false.
so you should use Identity operator in this case. following evaluations just returns true when a is null
(typeof a === "null")
// or
(a === null)
I am trying to use two predicates (say, methodsWiThSameParameters and methodsWiThSameReturn) from another one (i.e. checkOverriding) but i receive the following error: "There are no commands to execute". Any clues?
I also tried to use functions but with no success, either due to syntax or to functions do not return boolean values.
They are part of a java metamodel specified in Alloy, as i commented in some earlier questions.
pred checkOverriding[]{
//check accessibility of methods involved in overriding
no c1, c2: Class {
c1=c2.^extend
some m1, m2:Method |
m1 in c1.methods && m2 in c2.methods && m1.id = m2.id
&& methodsWiThSameParameters[m1, m2] && methodsWiThSameReturn[m1, m2] &&
( (m1.acc = protected && (m2.acc = private_ || #(m2.acc) = 0 )) ||
(m1.acc = public && (m2.acc != public || #(m2.acc) = 0 )) ||
(#(m1.acc) = 0 && m2.acc != private_ )
)
}
}
pred methodsWiThSameParameters [first,second:Method]{
m1.param=m2.param || (#(m1.param)=0 && #(m2.param)=0)
}
pred methodsWiThSameReturn [first, second:Method]{
m1.return=m2.return || (#(m1.return)=0 && #(m2.return)=0)
}
Thank you for your response, mr C. M. Sperberg-McQueen, but i think i was not clear enough in my question.
My predicate, say checkOverriding, is being called from a fact like this:
fact chackJavaWellFormednessRules{
checkOverriding[]
}
Thus, i continue not understanding the error: "There are no commands to execute" .
You've defined predicates; they have a purely declarative semantics and they will be true in some subset of instances of the model and false in the complementary subset.
If you want the Analyzer to do anything, you need to give it an instruction; the instruction to search for an instance of a predicate is run. So you'll want to say something like
run methodsWithSameParameters for 3
or
run methodsWithSameParameters for 5
run methodsWithSameReturn for 5
Note that you can have more than one instruction in an Alloy model; the Analyzer lets you tell it which to execute.
[Addendum]
The Alloy Analyzer regards the keywords run and check (and only them) as 'commands'. From your description, it sounds very much as if you don't have any occurrences of those keywords in the model.
If all you want to do is to see some instances of the Alloy model (to verify that the model is not self-contradictory), then the simplest way is to add something like the following to the model:
pred show {}
run show for 3
Or, if you already have a named predicate, you could simply add a run command for that predicate:
run checkOverriding
But without a clause in the model that begins with either run or check, you do not have a 'command' in the model.
You say that you have defined a predicate (checkOverriding) and then specified in a fact that that predicate is always satisfied. This amounts to saying that the predicate checkOverriding is always true (and might just as well be done by making checkOverriding a fact instead of a predicate), but it has a purely declarative meaning, and it does not count as a "command". If you want Alloy to find instances of a predicate, you must use the run command; if you want Alloy to find counter-examples for an assertion, you must use the check command.
At the beginning of a method I want to check if the method is called with these exact parameters before, and if so, return the result that was returned back then.
At first, with one parameter, I used a Dictionary, but now I need to check 3 parameters (a String, an Object and a boolean).
I tried making a custom Object like so:
var cacheKey:Object = { identifier:identifier, type:type, someBoolean:someBoolean };
//if key already exists, return it (not working)
if (resultCache[cacheKey]) return resultCache[cacheKey];
//else: create result ...
//and save it in the cache
resultCache[cacheKey] = result;
But this doesn't work, because the seccond time the function is called, the new cacheKey is not the same object as the first, even though it's properties are the same.
So my question is: is there a datatype that will check the properties of the object used as key for a matching key?
And what else is my best option? Create a cache for the keys as well? :/
Note there are two aspects to the technical solution: equality comparison and indexing.
The Cliff Notes version:
It's easy to do custom equality comparison
In order to perform indexing, you need to know more than whether one object is equal to another -- you need to know which is object is "bigger" than the other.
If all of your properties are primitives you should squash them into a single string and use an Object to keep track of them (NOT a Dictionary).
If you need to compare some of the individual properties for reference equality you're going to have a write a function to determine which set of properties is bigger than the other, and then make your own collection class that uses the output of the comparison function to implement its own a binary search tree based indexing.
If the number of unique sets of arguments is in the several hundreds or less AND you do need reference comparison for your Object argument, just use an Array and the some method to do a naive comparison to all cached keys. Only you know how expensive your actual method is, so it's up to you to decide what lookup cost (which depends on the number of unique arguments provided to the function) is acceptable.
Equality comparison
To address equality comparison it is easy enough to write some code to compare objects for the values of their properties, rather than for reference equality. The following function enforces strict set comparison, so that both objects must contain exactly the same properties (no additional properties on either object allowed) with the same values:
public static propsEqual(obj1:Object, obj2:Object):Boolean {
for(key1:* in obj1) {
if(obj2[key1] === undefined)
return false;
if(obj2[key1] != obj2[key1])
return false;
}
for(key2:* in obj2)
if(obj1[key2] === undefined)
return false;
return true;
}
You could speed it up by eliminating the second for loop with the tradeoff that {A:1, B:2} will be deemed equal to {A:1, B:2, C:'An extra property'}.
Indexing
The problem with this in your case is that you lose the indexing that a Dictionary provides for reference equality or that an Object provides for string keys. You would have to compare each new set of function arguments to the entire list of previously seen arguments, such as using Array.some. I use the field currentArgs and the method to avoid generating a new closure every time.
private var cachedArgs:Array = [];
private var currentArgs:Object;
function yourMethod(stringArg:String, objArg:Object, boolArg:Boolean):* {
currentArgs = { stringArg:stringArg, objArg:objArg, boolArg:boolArg };
var iveSeenThisBefore:Boolean = cachedArgs.some(compareToCurrent);
if(!iveSeenThisBefore)
cachedArgs.push(currentArgs);
}
function compareToCurrent(obj:Object):Boolean {
return someUtil.propsEqual(obj, currentArgs);
}
This means comparison will be O(n) time, where n is the ever increasing number of unique sets of function arguments.
If all the arguments to your function are primitive, see the very similar question In AS3, where do you draw the line between Dictionary and ArrayCollection?. The title doesn't sound very similar but the solution in the accepted answer (yes I wrote it) addresses the exact same techinical issue -- using multiple primitive values as a single compound key. The basic gist in your case would be:
private var cachedArgs:Object = {};
function yourMethod(stringArg:String, objArg:Object, boolArg:Boolean):* {
var argKey:String = stringArg + objArg.toString() + (boolArg ? 'T' : 'F');
if(cachedArgs[argKey] === undefined)
cachedArgs[argKey] = _yourMethod(stringArg, objArg, boolArg);
return cachedArgs[argKey];
}
private function _yourMethod(stringArg:String, objArg:Object, boolArg:Boolean):* {
// Do stuff
return something;
}
If you really need to determine which reference is "bigger" than another (as the Dictionary does internally) you're going to have to wade into some ugly stuff, since Adobe has not yet provided any API to retrieve the "value" / "address" of a reference. The best thing I've found so far is this interesting hack: How can I get an instance's "memory location" in ActionScript?. Without doing a bunch of performance tests I don't know if using this hack to compare references will kill the advantages gained by binary search tree indexnig. Naturally it would depend on the number of keys.
I have a question about if statement in tcl of the following code:
if {(($number == 1)&&($name == "hello")) || (($number == 0)&&($name == "yes"))} {
#do something here
}
The above code works, but if I wrote it down like this:
if {{$number == 1 && $name == "hello"} || {$number == 0&&$name == "yes"}} {
#do something here
}
It complains that the $number is expected to be a boolean, why? Is the second one not a valid expression? How to correct it?
Braces, {}, and parentheses, (), are not interchangeable in Tcl.
Formally, braces are (with one exception) a kind of quoting that indicates that no further substitutions are to be performed on the contents. In the first case above, this means that that argument is delivered to if without substitution, which evaluates it as an expression. The expression sub-language has a strongly-analogous brace interpretation scheme to general Tcl; they denote a literal value with no further substitutions to perform on it.
By contrast, parentheses are mostly not special in Tcl. The exceptions are in the names of elements of arrays (e.g., $foo(bar)), in the expression sublanguage (which uses them for grouping, as in mathematical expressions all over programming) and in the regular expression sublanguage (where they are a different type of grouping and a few other things). It is entirely legal to use parentheses — balanced or otherwise — as part of a command name in Tcl, but you might have your fellow programmers complaining at you anyway for writing confusing code.
The Specifics
In this specific case, the test expression of this if:
if {{$number == 1 && $name == "hello"} || {$number == 0&&$name == "yes"}} {...}
is parsed into:
blah#1 LOGICAL_OR blah#2
where each blah is a literal. Unfortunately, blah#1 (which is exactly equal to $number == 1 && $name == "hello") has no boolean interpretation. (Nor does blah#2 but we never bother to consider that.) Things are definitely going very wrong here!
The simplest fix is to change those bogus braces back to parentheses:
if {($number == 1 && $name == "hello") || ($number == 0&&$name == "yes")} {...}
I bet this is what you originally wanted.
Warning: Advanced Topic
However, the other fix is to add in a little extra:
if {[expr {$number == 1 && $name == "hello"}] || [expr {$number == 0&&$name == "yes"}]} {...}
This is normally not a good idea — extra bulk for no extra gain — but it makes sense where you're trying to use a dynamically-generated expression as a test condition. Do not do this unless you're really really certain you need to do this! I mean it. It's a very advanced technique that you hardly ever need, and there's often a better way to do your overall goal. If you think you might need it, for goodness' sake ask here on SO and we'll try to find a better way; there's almost always one available.
I think the error message you are getting does not mean that $number has to be a boolean (I got the message expected boolean value but got "$number == 1 && $name == "hello"").
It means that the string $number == 1 && $name == "hello" is not a boolean value - which is definitely true.
If you use curly braces in your if expression those strings are not evaluated but are simply interpreted as they are - as a string of characters.
In short: if uses a special "mini language" for its condition script — the same understood by the expr command. This is stated in the if manual page:
The if command evaluates expr1 as an expression (in the same way that expr evaluates its argument).
In contrast to Tcl itself, which is quite LISP-like and/or Unix shell-like, that "expr mini language" is way more "traditional" in the sense it feels like C.
In ($number == 1) number is assigned 1 and the comparison is made.Ex: 1==1 here output is Boolean.
But in {$number == 1 && $name == "hello"} $number is not assigned because of flower bracket $number is compared with 1 so output obtained is not Boolean.
I keep seeing code that does checks like this
if (IsGood == false)
{
DoSomething();
}
or this
if (IsGood == true)
{
DoSomething();
}
I hate this syntax, and always use the following syntax.
if (IsGood)
{
DoSomething();
}
or
if (!IsGood)
{
DoSomething();
}
Is there any reason to use '== true' or '== false'?
Is it a readability thing? Do people just not understand Boolean variables?
Also, is there any performance difference between the two?
I follow the same syntax as you, it's less verbose.
People (more beginner) prefer to use == true just to be sure that it's what they want. They are used to use operator in their conditional... they found it more readable. But once you got more advanced, you found it irritating because it's too verbose.
I always chuckle (or throw something at someone, depending on my mood) when I come across
if (someBoolean == true) { /* ... */ }
because surely if you can't rely on the fact that your comparison returns a boolean, then you can't rely on comparing the result to true either, so the code should become
if ((someBoolean == true) == true) { /* ... */ }
but, of course, this should really be
if (((someBoolean == true) == true) == true) { /* ... */ }
but, of course ...
(ah, compilation failed. Back to work.)
I would prefer shorter variant. But sometimes == false helps to make your code even shorter:
For real-life scenario in projects using C# 2.0 I see only one good reason to do this: bool? type. Three-state bool? is useful and it is easy to check one of its possible values this way.
Actually you can't use (!IsGood) if IsGood is bool?. But writing (IsGood.HasValue && IsGood.Value) is worse than (IsGood == true).
Play with this sample to get idea:
bool? value = true; // try false and null too
if (value == true)
{
Console.WriteLine("value is true");
}
else if (value == false)
{
Console.WriteLine("value is false");
}
else
{
Console.WriteLine("value is null");
}
There is one more case I've just discovered where if (!IsGood) { ... } is not the same as if (IsGood == false) { ... }. But this one is not realistic ;) Operator overloading may kind of help here :) (and operator true/false that AFAIK is discouraged in C# 2.0 because it is intended purpose is to provide bool?-like behavior for user-defined type and now you can get it with standard type!)
using System;
namespace BoolHack
{
class Program
{
public struct CrazyBool
{
private readonly bool value;
public CrazyBool(bool value)
{
this.value = value;
}
// Just to make nice init possible ;)
public static implicit operator CrazyBool(bool value)
{
return new CrazyBool(value);
}
public static bool operator==(CrazyBool crazyBool, bool value)
{
return crazyBool.value == value;
}
public static bool operator!=(CrazyBool crazyBool, bool value)
{
return crazyBool.value != value;
}
#region Twisted logic!
public static bool operator true(CrazyBool crazyBool)
{
return !crazyBool.value;
}
public static bool operator false(CrazyBool crazyBool)
{
return crazyBool.value;
}
#endregion Twisted logic!
}
static void Main()
{
CrazyBool IsGood = false;
if (IsGood)
{
if (IsGood == false)
{
Console.WriteLine("Now you should understand why those type is called CrazyBool!");
}
}
}
}
}
So... please, use operator overloading with caution :(
According to Code Complete a book Jeff got his name from and holds in high regards the following is the way you should treat booleans.
if (IsGood)
if (!IsGood)
I use to go with actually comparing the booleans, but I figured why add an extra step to the process and treat booleans as second rate types. In my view a comparison returns a boolean and a boolean type is already a boolean so why no just use the boolean.
Really what the debate comes down to is using good names for your booleans. Like you did above I always phrase my boolean objects in the for of a question. Such as
IsGood
HasValue
etc.
The technique of testing specifically against true or false is definitely bad practice if the variable in question is really supposed to be used as a boolean value (even if its type is not boolean) - especially in C/C++. Testing against true can (and probably will) lead to subtle bugs:
These apparently similar tests give opposite results:
// needs C++ to get true/false keywords
// or needs macros (or something) defining true/false appropriately
int main( int argc, char* argv[])
{
int isGood = -1;
if (isGood == true) {
printf( "isGood == true\n");
}
else {
printf( "isGood != true\n");
}
if (isGood) {
printf( "isGood is true\n");
}
else {
printf( "isGood is not true\n");
}
return 0;
}
This displays the following result:
isGood != true
isGood is true
If you feel the need to test variable that is used as a boolean flag against true/false (which shouldn't be done in my opinion), you should use the idiom of always testing against false because false can have only one value (0) while a true can have multiple possible values (anything other than 0):
if (isGood != false) ... // instead of using if (isGood == true)
Some people will have the opinion that this is a flaw in C/C++, and that may be true. But it's a fact of life in those languages (and probably many others) so I would stick to the short idiom, even in languages like C# that do not allow you to use an integral value as a boolean.
See this SO question for an example of where this problem actually bit someone...
isalpha() == true evaluates to false??
I agree with you (and am also annoyed by it). I think it's just a slight misunderstanding that IsGood == true evaluates to bool, which is what IsGood was to begin with.
I often see these near instances of SomeStringObject.ToString().
That said, in languages that play looser with types, this might be justified. But not in C#.
Some people find the explicit check against a known value to be more readable, as you can infer the variable type by reading. I'm agnostic as to whether one is better that the other. They both work. I find that if the variable inherently holds an "inverse" then I seem to gravitate toward checking against a value:
if(IsGood) DoSomething();
or
if(IsBad == false) DoSomething();
instead of
if(!IsBad) DoSomething();
But again, It doen't matter much to me, and I'm sure it ends up as the same IL.
Readability only..
If anything the way you prefer is more efficient when compiled into machine code. However I expect they produce exactly the same machine code.
From the answers so far, this seems to be the consensus:
The short form is best in most cases. (IsGood and !IsGood)
Boolean variables should be written as a positive. (IsGood instead of IsBad)
Since most compilers will output the same code either way, there is no performance difference, except in the case of interpreted languages.
This issue has no clear winner could probably be seen as a battle in the religious war of coding style.
I prefer to use:
if (IsGood)
{
DoSomething();
}
and
if (IsGood == false)
{
DoSomething();
}
as I find this more readable - the ! is just too easy to miss (in both reading and typing); also "if not IsGood then..." just doesn't sound right when I hear it, as opposed to "if IsGood is false then...", which sounds better.
It's possible (although unlikely, at least I hope) that in C code TRUE and FALSE are #defined to things other than 1 and 0. For example, a programmer might have decided to use 0 as "true" and -1 as "false" in a particular API. The same is true of legacy C++ code, since "true" and "false" were not always C++ keywords, particularly back in the day before there was an ANSI standard.
It's also worth pointing out that some languages--particularly script-y ones like Perl, JavaScript, and PHP--can have funny interpretations of what values count as true and what values count as false. It's possible (although, again, unlikely on hopes) that "foo == false" means something subtly different from "!foo". This question is tagged "language agnostic", and a language can define the == operator to not work in ways compatible with the ! operator.
I've seen the following as a C/C++ style requirement.
if ( true == FunctionCall()) {
// stuff
}
The reasoning was if you accidentally put "=" instead of "==", the compiler will bail on assigning a value to a constant. In the meantime it hurts the readability of every single if statement.
Occasionally it has uses in terms of readability. Sometimes a named variable or function call can end up being a double-negative which can be confusing, and making the expected test explicit like this can aid readability.
A good example of this might be strcmp() C/C++ which returns 0 if strings are equal, otherwise < or > 0, depending on where the difference is. So you will often see:
if(strcmp(string1, string2)==0) { /*do something*/ }
Generally however I'd agree with you that
if(!isCached)
{
Cache(thing);
}
is clearer to read.
I prefer the !IsGood approach, and I think most people coming from a c-style language background will prefer it as well. I'm only guessing here, but I think that most people that write IsGood == False come from a more verbose language background like Visual Basic.
Only thing worse is
if (true == IsGood) {....
Never understood the thought behind that method.
The !IsGood pattern is eaiser to find than IsGood == false when reduced to a regular expression.
/\b!IsGood\b/
vs
/\bIsGood\s*==\s*false\b/
/\bIsGood\s*!=\s*true\b/
/\bIsGood\s*(?:==\s*false|!=\s*true)\b/
For readability, you might consider a property that relies on the other property:
public bool IsBad => !IsGood;
Then, you can really get across the meaning:
if (IsBad)
{
...
}
I prefer !IsGood because to me, it is more clear and consise. Checking if a boolean == true is redundant though, so I would avoid that. Syntactically though, I don't think there is a difference checking if IsGood == false.
In many languages, the difference is that in one case, you are having the compiler/interpreter dictate the meaning of true or false, while in the other case, it is being defined by the code. C is a good example of this.
if (something) ...
In the above example, "something" is compared to the compiler's definition of "true." Usually this means "not zero."
if (something == true) ...
In the above example, "something" is compared to "true." Both the type of "true" (and therefor the comparability) and the value of "true" may or may not be defined by the language and/or the compiler/interpreter.
Often the two are not the same.
You forgot:
if(IsGood == FileNotFound)
It seems to me (though I have no proof to back this up) that people who start out in C#/java type languages prefer the "if (CheckSomething())" method, while people who start in other languages (C++: specifically Win32 C++) tend to use the other method out of habit: in Win32 "if (CheckSomething())" won't work if CheckSomething returns a BOOL (instead of a bool); and in many cases, API functions explicitly return a 0/1 int/INT rather than a true/false value (which is what a BOOL is).
I've always used the more verbose method, again, out of habit. They're syntactically the same; I don't buy the "verbosity irritates me" nonsense, because the programmer is not the one that needs to be impressed by the code (the computer does). And, in the real world, the skill level of any given person looking at the code I've written will vary, and I don't have the time or inclination to explain the peculiarities of statement evaluation to someone who may not understand little unimportant bits like that.
Ah, I have some co-worked favoring the longer form, arguing it is more readable than the tiny !
I started to "fix" that, since booleans are self sufficient, then I dropped the crusade... ^_^ They don't like clean up of code here, anyway, arguing it makes integration between branches difficult (that's true, but then you live forever with bad looking code...).
If you write correctly your boolean variable name, it should read naturally:
if (isSuccessful) vs. if (returnCode)
I might indulge in boolean comparison in some cases, like:
if (PropertyProvider.getBooleanProperty(SOME_SETTING, true) == true) because it reads less "naturally".
For some reason I've always liked
if (IsGood)
more than
if (!IsBad)
and that's why I kind of like Ruby's unless (but it's a little too easy to abuse):
unless (IsBad)
and even more if used like this:
raise InvalidColor unless AllowedColors.include?(color)
Cybis, when coding in C++ you can also use the not keyword. It's part of the standard since long time ago, so this code is perfectly valid:
if (not foo ())
bar ();
Edit: BTW, I forgot to mention that the standard also defines other boolean keywords such as and (&&), bitand (&), or (||), bitor (|), xor (^)... They are called operator synonyms.
If you really think you need:
if (Flag == true)
then since the conditional expression is itself boolean you probably want to expand it to:
if ((Flag == true) == true)
and so on. How many more nails does this coffin need?
If you happen to be working in perl you have the option of
unless($isGood)
I do not use == but sometime I use != because it's more clear in my mind. BUT at my job we do not use != or ==. We try to get a name that is significat if with hasXYZ() or isABC().
Personally, I prefer the form that Uncle Bob talks about in Clean Code:
(...)
if (ShouldDoSomething())
{
DoSomething();
}
(...)
bool ShouldDoSomething()
{
return IsGood;
}
where conditionals, except the most trivial ones, are put in predicate functions. Then it matters less how readable the implementation of the boolean expression is.
We tend to do the following here:
if(IsGood)
or
if(IsGood == false)
The reason for this is because we've got some legacy code written by a guy that is no longer here (in Delphi) that looks like:
if not IsNotGuam then
This has caused us much pain in the past, so it was decided that we would always try to check for the positive; if that wasn't possible, then compare the negative to false.
The only time I can think where the more vebose code made sense was in pre-.NET Visual Basic where true and false were actually integers (true=-1, false=0) and boolean expressions were considered false if they evaluated to zero and true for any other nonzero values. So, in the case of old VB, the two methods listed were not actually equivalent and if you only wanted something to be true if it evaluated to -1, you had to explicitly compare to 'true'. So, an expression that evaluates to "+1" would be true if evaluated as integer (because it is not zero) but it would not be equivalent to 'true'. I don't know why VB was designed that way, but I see a lot of boolean expressions comparing variables to true and false in old VB code.