Where can i find a concrete document or a dos and donts documentation on using C++ with Objective-C?
Apple seems to have removed that document from their website and i am all puzzled with collating bits of information from blogs and questions posted here.
Anyone can guide about the same.
When do we use .mm file, while mixing syntax or while using an object in .m file which belongs to a C++ class ?
While passing objects between functions belonging to two different language like passing objective-c object to a function in cpp file is it necessary to collect it in void * or can I use (ObjectiveC inteface)*?
You need to use Objective-C++ whenever you are either #include/#importing or directly writing both Objective-C and C++ code in the same file. It's usually obvious with explicit code; the #includes are often less so, and you need to take care to avoid "leaking" one of the languages into too much of the other. Your example of #importing a C++ header file is clear-cut: you can only do that with Objective-C++. Note that if your Cplusplus was a struct type, you could forward-declare it as such instead of #importing a C++ header.
If you do this in a header, it will "poison" the header to only work in that mode. You'll need to actively avoid this, or your whole project will soon end up with only .mm files. I have documented some techniques in this article and previously, in this earlier article. For newer versions of Objective-C, you can also add ivars to classes in category extensions. This means you can define C++-typed ivars in your .mm file, not the header, which allows .m files to #import it.
For your second question (Please only ask one question at a time): the id type is defined in the objc.h header file in terms of C and thus will work in both Objective-C and C++. Likewise, the Objective-C runtime API is exposed in terms of C functions, which will work from C++, too. If you actually want to send messages and access properties on Objective-C objects with the canonical syntax from C++ code, you'll need to switch that file to Objective-C++.
Use .mm files when you have a c++ syntax in your code or when including file(s) which contain c++ code.
Mixing C++ with objective-c may be a bit confusing but if you think pointer-wise than it is not a big deal. Treat C++ object instance methods as you would in C++ and the same goes for objective c objects.
Related
i am currently developing a simple app on my own.
I want to use a library from some sample app. The library consists of .mm files (C++ code) and runs perfectly on the original project.
However, when i copy the library to my file (Objective C), it seems that my other object files do not see the file. All the classes and protocol defined in this library are not recognised at all.
Errors are such as "No type or protocol name", "Unknown typename "classname" " when i try to use the library classes and protocols.
I have spent quite a lot time searching, but to no veil.
Thanks in advance.
Objective-C simply won't understand the C++ language constructs like class, etc. If you want to use this Objective-C++ code in your app then your app also needs to become Objective-C++ which can be done simply by renaming all your source files from *.m to *.mm.
In the case of Swift, however, you cannot expose C++ to it at all, and can only integrate it with C or Objective-C, so you are therefore forced to create a pure Objective-C wrapper for the C++ code (i.e. implementation in .mm but exposing no C++ types in its header file).
As I understand correctly, besides the fact that Objective-C language is a strict superset of a "clean" C, added OOP paradigm is simulated by a set of functions partially described in Objective-C Runtime Reference.
Therefore, I'm expecting a possibility to somehow compile Objective-C code in an intermediate C/C++ file (maybe with some asm inserts).
Is it generally possible ?
You could use the clang rewriter to convert to C++. Not aware of a way to go to C though.
The rewriter is available via the "-rewrite-objc" command line option.
As far as I know, there is no software that preprocesses Objective-C code into intermediate C code.
But you could write your Objective-C program entirely in C by calling directly into the Objective-C runtime. The trouble is just that the code might vary between implementations or even different versions of the same runtime.
The question is, is it actually worth the trouble?
I want to create an Objective C class at runtime from a file. So for example I have an objective c application, I want to then point it at a text file (or an .h .m pair, whatever works) and then have the application parse the text file and create the class at runtime (Class no object). Of course I would write the parser and all of that stuff I just want to know if its possible. I read these two articles:
http://www.mikeash.com/pyblog/friday-qa-2010-11-6-creating-classes-at-runtime-in-objective-c.html
http://www.mikeash.com/pyblog/friday-qa-2010-11-19-creating-classes-at-runtime-for-fun-and-profit.html
Which shows how to make an objective C class at runtime, but its being done using C functions which were defined at compile time. If I can find a way to do the same thing using strings to define the functions that would be perfect, because then I don't have to define them at compile time.
That's what called reflective programming. I guess there's no code evaluation support for Obj-C since it's not a scripting language, so the reflection concept for Obj-C is quietly limited. Plus, at run-time the compiler already translate the code into Obj-C clang code and it's a very time-consuming job just to reverse-translate the bytecode and recompile it again
For Obj-C reflection you can refer to these answers
Build a class :
Create objective-c class instance by name?
Implement a method :
Objective-C, how can i hook up a method in another class
Change class for an object :
Objective-C: How to change the class of an object at runtime?
Sure. Totally possible.
I would suggest starting with the Objective-C support in this as it includes a full-on Objective-C parser and code generator.
see my github project cocoa-interprreter it does part of what you want.
it takes a text file and compiles it at runtime .. and then runs the resulting executable using NSTask. It would be quite easy to change it so the binary is loaded into the own process using NSBundle
https://github.com/Daij-Djan/cocoa-interpreter
Baby new to Xcode, Cocoa touch and iOS development in general. And am taking the Stanford walkthrough for their iPhone class. I am a little confused at a couple of places and need to shoot my doubts to you guys:
I have two classes that I have created for my model, essentially CalculatorBrain.m and CalculatorBrain.h.
From what I gather, in Objective C, creating a class essentially consists of two functions, one is to declare the class which contains the method/messages and other variables while the other is the actual implementation for the same. From this stems two questions:
Why must I declare a class without implementing it's methods at first? (the concept seems to be borrowed from interfaces) and only then move on to implementing it .
From the above question, as I go through the walkthrough, I notice that the class declaration took place in CalculatorBrain.h whereas the methods were actually implemented in CalculatorBrain.m. I am unable to grasp the nuances of why this was done so if anybody is willing to shed some light on this, it would really help
Thanks again,
Parijat Kalia!!!
These are traditions from the C world, and they're just common practice to avoid some problems. They aren't two classes, they are the definition (in the .h file) and the implementation (in the .c or .cc file).
If you defined the class in the .c file, you couldn't refer to it elsewhere because it wouldn't be defined. You could include your .c file, but then you'd have two copies of the code. You could also use the "extern" keyword, but at this point it's kind of odd.
If you put code in the .h file, then when it's included the code gets included. This means you can get compiler errors that you have three "getMyThing" functions.
This means you can give out your headers to others without giving away your top-secret implementation (useful for making libraries), include your header without worrying about the possible multiple definitions, etc. You can also add variables and functions in the .c file which people using the header (like your other code) can't see or use, so you don't have to worry about changing it later and having compilation break.
Or in more practical terms: If I compile a file containing just Obj-C code or just C++ code under Objective-C++, can I be sure they'll behave exactly as they did before? In what ways might they differ?
More or less, that is true; anything that compiles under either C++ or Objective-C will compile with the Objective-C++ compiler.
However, there are some edges where this doesn't hold true and a handful of behavioral differences.
I'm not aware of any specific at-runtime behavioral differences, though. The differences should generally show up during compilation and mostly as problems caused by C++'s "enhanced" notion of types.
Operator overloading can be exceedingly problematic, however. If a body of code makes extensive use of operator overloading, that can cause issue when mixing Objective-C in for the first time. But, again, typically as compiler warnings and rarely as at-runtime bugs.
Objective-C is a superset of C only. You can mix C++ with Objective-C - that is called Objective-C++ - however there are certain rules that you have to follow.
Here you can see all the rules that have to be followed: http://developer.apple.com/library/mac/#documentation/Cocoa/Conceptual/ObjectiveC/Articles/ocCPlusPlus.html
EDIT: This is a quote from the developer website: "Apple’s Objective-C compiler allows you to freely mix C++ and Objective-C code in the same source file. This Objective-C/C++ language hybrid is called Objective-C++. With it you can make use of existing C++ libraries from your Objective-C applications.".
So it clearly says, that the mixture of C++ and Objective-C is called Objective-C++, but that is not a superset of C++, but it's two languages that with some rules can be used together. (For instance you can't mix objective-c and c++ classes during inheritance, but there are others as well.)
I guess there's at least one example of valid Objective-C code that's not valid Objective-C++:
void SayHello()
{
char* this = "hello!";
printf("%s", this);
}
(what someone would be doing naming a variable 'this', I don't know)