This relates to the Dolphin variant of Smalltalk.
I'm digging around in the image to try and figure this out but haven't
find the correct method invocation yet and I'm hoping someone might be
able to help shortcut this process. What I'm trying to do is to find
all methods (either within the entire system or, preferably, just
within a single class) which refer to a given string, symbol, or
method. I've found the #references family of methods in
SmalltalkSystem but have not had luck figuring out how to get them to
give back something resembling what I want.
The programmatic way, here we go
SmalltalkSystem current browseContainingText: 'Dolphin'.
I don't have Dolphin at hand, but the following code should work in all Smalltalk with the refactoring engine (this includes Dolphin):
result := BrowserEnvironment new matches: 'Dolphin'.
Then you can iterate over the results like this:
result classesAndSelectorsDo: [ :class :selector | ... ].
Related
Let's say I have a type of object in my game called oCharacter. All characters must have names, so I want to provide one when I construct the object. I can do that by using the _variables argument of instance_create_layer:
instance_create_layer(0, 0, "Instances", oCharacter, { name: "George" });
I could even make sure that I don't forget to do this by making a "constructor" function for characters and only instantiating them using that:
function character_create(_x, _y, _name) {
return instance_create_layer(_x, _y, "Instances", oCharacter, { name: _name });
}
But this approach has two problems.
The first is that I or another developer might forget about this convention and instantiate a character directly using instance_create_layer, forgetting to pass a name and setting up a runtime error further down the road.
The second (related) issue is that Feather doesn't know about this convention, so my Feather window is full of error messages about references to instance variables that aren't declared in the Create event - but I don't see how I can declare these variables in the Create event, as I'm expecting their value to be provided by the creator.
Is there some way of doing this that addresses these issues?
The first problem is just about setting rules about the code conventions within your team, if your team does not know about these conventions you want them to follow, then you should tell it them in a meeting.
For the second problem: Maybe you could create an empty/nullable variable in the Create Event? I'm afraid I'm not familiar with Feather
Personally I would do two things for this.
Create development standards for the team and put them in something like a Word document, wiki page, onenote, whatever makes the most sense for your team.
I would use a function to create the instance of the object (like you're doing there), and have some simple validation checks inside of the create event itself that will cancel it's creation (something like a guard clause) and output a debug message with a reminder.
It's not the most elegant solution but that should do the trick (assuming you haven't found something else by now haha)
One common problem we have in our codebase is that people forget to check if a list is empty before using it in an in clause.
For example (in Scala with Anorm):
def exists(element: String, list: List[String]): Boolean =
SQL("select {element} in {list} as result")
.on('element -> element, 'list -> list)
.as(SqlParser.bool("result").single)
This code works perfectly well as long as list has at least one element.
If it has 0 elements, you get a syntax error, which is weird if you're used to other programming languages that would allow this empty list case.
So, my question is: what's the best way to prevent this error from happening?
Initially, we did this:
def exists(element: String, list: List[String]): Boolean =
if (list.nonEmpty) {
SQL("select {element} in {list} as result")
.on('element -> element, 'list -> list)
.as(SqlParser.bool("result").single)
} else {
false
}
This works perfectly well, and has the added advantage that it doesn't hit the database at all.
Unfortunately, we don't remember to do this every time, and it seems that 1-2 times a month we're fixing an issue related to this.
An alternate solution we came up with was to use a NonEmptyList class instead of a standard List. This class must have at least one element. This works excellent, but again, people have not been diligent with always using this class.
So I'm wondering if there's an approach I'm missing that prevent this type of error better?
It looks like you've already found a way to resolve this problem - you have an exists() function which handles an empty list cleanly. The problem is that people are writing their own exists() functions which don't do that.
You need to make sure that your function is accessible as a utility function, so that you can reuse it whenever you need to, rather than having to rewrite the function.
Your problem is an encapsulation problem: the Anorm API is like an open flame and people can burn themselves. If you rely just on people to take precautions, someone will get burnt.
The solution is to restrict the access to the Anorm API to a limited module/package/area of your code:
Anorm API will be private and accessible only from very few places, where it is going to be easy to perform the necessary controls. This part of the code will expose an API
Every other part of the code will need to go through that API, effectively using Anorm in the "safe" way
I'm currently working on my first major project in clojure and have run into a question regarding coding style and the most "clojure-esque" way of doing something. Basically I have a function I'm writing which takes in a data structure and a template that the function will try to massage the data structure into. The template structure will look something like this:
{
:key1 (:string (:opt :param))
:key2 (:int (:opt :param))
:key3 (:obj (:tpl :template-structure))
:key4 (:list (:tpl :template-structure))
}
Each key is an atom that will be searched for in the given data structure, and it's value will be attempted to be matched to the type given in the template structure. So it would look for :key1 and check that it's a string, for instance. The return value would be a map that has :key1 pointing to the value from the given data structure (the function could potentially change the value depending on the options given).
In the case of :obj it takes in another template structure, and recursively calls itself on that value and the template structure, and places the result from that in the return. However, if there's an error I want that error returned directly.
Similarly for lists I want it to basically do a map of the function again, except in the case of an error which I want returned directly.
My question is what is the best way to handle these errors? Some simple exception handling would be the easiest way, but I feel that it's not the most functional way. I could try and babysit the errors all the way up the chain with tons of if statements, but that also doesn't seem very sporting. Is there something simple I'm missing or is this just an ugly problem?
You might be interested in schematic, which does pretty similar stuff. You can see how it's used in the tests, and the implementation.
Basically I defined an error function, which returns nil for correctly-formatted data, or a string describing the error. Doing it with exceptions instead would make the plumbing easier, but would make it harder to get the detailed error messages like "[person: [first_name: expected string, got integer]]".
How do I list all unimplemented methods in a package? Given that the method should be implemented in that package and not in other (for example in a superclass outside the package or in Object).
Edit: Yes, I'd like to know "messages sent that are not implemented" but to limit the analysis to one specific given package, for ex:
FooPackage unimplementedMessageSends.
I haven't fully tested this so you might need to tweak it a bit to suit your requirements, but from how I understand your question, you might be after something like this:
| allMethodsSent allMethodsImplemented |
allMethodsSent := IdentitySet new.
allMethodsImplemented := IdentitySet new.
(SystemOrganization listAtCategoryNamed: #'Collections-Arrayed')
do: [:eachClassName |
(Smalltalk at: eachClassName) methodDictionary valuesDo: [:eachMethod |
allMethodsSent addAll: eachMethod messages.
].
allMethodsImplemented addAll: (Smalltalk at: eachClassName) selectors
].
^allMethodsSent
removeAllFoundIn: allMethodsImplemented;
yourself
Hopefully that'll help you get started at least, if you need to tweak it, have a look at the classes Behavior (to see what you can use, not to change it!), CompiledMethod, and SystemOrganization.
Obviously this example uses the category (I'm assuming that's what you mean by package?) "Collections-Arrayed", but you could of course adapt it to make that a method parameter, something like:
MyUtilClass unimplementedMessageSendsFor: aCategorySymbol
I'm trying to store the names of some variables inside strings. For example:
Dim Foo1 as Integer
Dim Foo1Name as String
' -- Do something to set Foo1Name to the name of the other variable --
MessageBox.Show(Foo1Name & " is the variable you are looking for.")
' Outputs:
' Foo1 is the variable you are looking for.
This would help with some debugging I'm working on.
Well, you can clearly just set Foo1Name = "Foo1" - but I strongly suspect that's not what you're after.
How would you know which variable you're trying to find the name of? What's the bigger picture? What you want may be possible with reflection, if we're talking about non-local variables, but I suspect it's either not feasible, or there's a better way to attack the problem in the first place.
Does this example from msdn using reflection help?
One solution would be to use an associative array to store your variables. Once, I did this in .Net, but I think I wrote a custom class to do it.
myArray("foo1Name") = "foo1"
Then, you can just store a list of your variable names, or you can wrap that up in the same class.
if( myArray(variableName(x)) == whatImLookingFor ) print variableName(x) & "is it"
I think this really depends on what you are trying to debug. Two possible things to look at are the Reflection and StackTrace classes. That said when your program is compiled, the compiler and runtime do not guarantee that that names need to be consistent with the original program.
This is especially the case with debug vs. release builds. The point of the .PDB files (symbols) in the debug version are to include more information about the original program. For native C/C++ applications it is strongly recommended that you generate symbols for every build (debug+release) of your application to help with debugging. In .NET this is less of an issue since there are features like Reflection. IIRC John Robbins recommends that you always generate symbols for .NET projects too.
You might also find Mike Stall's blog useful and the managed debugger samples.
For finding the variable name, see: Finding the variable name passed to a function
This would apply to VB.Net as well.