Why we need to alloc , init for a mutable object

Question

I don't know if the question was asked previously and Im searching for some good answer.

Question is : Whenever sometimes I don't use [[... alloc] init] for some mutable class , I get crash.

Example :

NSMutableDictionary *myDict = someObject ; //[allocation of some other dictionary object directly without using alloc , init].

For some stages compiler warns me at runtime that you can not change values within myDict even though it is mutable.

Instead of that if I write code :

NSMutableDictionary *myDict = [[NSMutableDictionary alloc] initWithDictionary:someObject]; 

then it works.

So why alloc , init is necessary ? Or What is the problem actually ?

Thanks.

Answer 1

Objective-C is based on C that is not a strongly typed language : you can put object of some type in variable of other type (basically because, in objective-C, they are all represented by their pointers).

When you write:

NSDictionary *immutableDict = [[NSDictionary alloc] init...];
NSMutableDictionary *mutableDict = immutableDict;

you just store the pointer to the immutable dictionary into the variable prepared to store a pointer to a mutable dictionary. But it doesn't magically transform your immutable dictionary in a mutable one.

But when you write:

NSMutableDictionary *mutableDict = [[NSMutableDictionary alloc] initWithDictionary:immutableDict];

you create a mutable dictionary that you initialize (fill) with the values contained in your immutable dictionary. So you can now modifiy your newly allocated mutable dictionary because it is mutable.

Answer 2

This is very basic question there are already number of solutions present to your question ( to understand alloc, init, new refer following )

alloc, init, and new in Objective-C

Use of alloc init instead of new

alloc and init what do they actually do