brief introduction

When we have more and more functions, the volume of a single file becomes larger, even bloated. When many people cooperate to develop a document, there will be many conflicts. At this time, we need to use the corresponding solution, that is category.
This article will proceed from several directions to understand the implementation mechanism of category in detail.

1. Introduction to category.
2. The difference between category and extension
 3. How does category load and add to a class
 4. category object association

Introduction to category

category is a feature added after 2.0. Its function is to add methods to existing files. We list several common scenarios:

1. Simulated multiple inheritance
 2. Publish private API s
 3. Decentralize the code in a class to manage it
 4. Private methods
5. ......

category has a lot to dig out, so let's unveil it step by step.

The Difference between category and extension

extension is determined at compile time and category is determined at run time, which is the fundamental difference between them. It determines the division and distinction between them.

The life cycle of extension follows the main class to hide private information. You must have the implementation / source code of this class before you can add extension to it. see extension

category can't add instance variables. During the run time, the object memory layout has been confirmed. At this time, you can't destroy the existing memory space, so you can't add instance variables.

How does category load and add to a class

In the runtime layer of OC, they are all represented by struct, and category is represented by category_t (which can be found in objc-runtime-new.h).

typedef struct category_t {
    const char *name;  //Class name
    classref_t cls;  //class
    struct method_list_t *instanceMethods;  //List of added instance methods
    struct method_list_t *classMethods;  //List of added class methods
    struct protocol_list_t *protocols;  //Additional protocol list
    struct property_list_t *instanceProperties;  //All attributes added
} category_t;

So how does category load?
In the OC runtime, the entry method is as follows (in objc-os.mm)

void _objc_init(void)
{
    static bool initialized = false;
    if (initialized) return;
    initialized = true;

    // fixme defer initialization until an objc-using image is found?
    environ_init();
    tls_init();
    lock_init();
    exception_init();

    // Register for unmap first, in case some +load unmaps something
    _dyld_register_func_for_remove_image(&unmap_image);
    dyld_register_image_state_change_handler(dyld_image_state_bound,
                                             1/*batch*/, &map_images);
    dyld_register_image_state_change_handler(dyld_image_state_dependents_initialized, 0/*not batch*/, &load_images);
}

category is attached to the class and sent when map_images. Under the new-ABI standard, map_images will eventually call the _read_images method in the objc-runtime-new.mm file. Let's see:

void _read_images(header_info **hList, uint32_t hCount)
{
    ...
        _free_internal(resolvedFutureClasses);
    }

    // Discover categories. 
    for (EACH_HEADER) {
        category_t **catlist =
            _getObjc2CategoryList(hi, &count);
        for (i = 0; i < count; i++) {
            category_t *cat = catlist[i];
            Class cls = remapClass(cat->cls);

            if (!cls) {
                // Category's target class is missing (probably weak-linked).
                // Disavow any knowledge of this category.
                catlist[i] = nil;
                if (PrintConnecting) {
                    _objc_inform("CLASS: IGNORING category \?\?\?(%s) %p with "
                                 "missing weak-linked target class",
                                 cat->name, cat);
                }
                continue;
            }

            // Process this category. 
            // First, register the category with its target class. 
            // Then, rebuild the class's method lists (etc) if 
            // the class is realized. 
            BOOL classExists = NO;
            if (cat->instanceMethods ||  cat->protocols
                ||  cat->instanceProperties)
            {
                addUnattachedCategoryForClass(cat, cls, hi);
                if (cls->isRealized()) {
                    remethodizeClass(cls);
                    classExists = YES;
                }
                if (PrintConnecting) {
                    _objc_inform("CLASS: found category -%s(%s) %s",
                                 cls->nameForLogging(), cat->name,
                                 classExists ? "on existing class" : "");
                }
            }

            if (cat->classMethods  ||  cat->protocols
                /* ||  cat->classProperties */)
            {
                addUnattachedCategoryForClass(cat, cls->ISA(), hi);
                if (cls->ISA()->isRealized()) {
                    remethodizeClass(cls->ISA());
                }
                if (PrintConnecting) {
                    _objc_inform("CLASS: found category +%s(%s)",
                                 cls->nameForLogging(), cat->name);
                }
            }
        }
    }

    // Category discovery MUST BE LAST to avoid potential races 
    // when other threads call the new category code before 
    // this thread finishes its fixups.

    // +load handled by prepare_load_methods()

    ...
}

Here we can see that

1. Add instance methods and attributes of category to the main class
2. Adding category's class methods to metaclasses

In either case, the class data is finally reorganized by calling the static void remethodizeClass(Class cls) function.

static void remethodizeClass(class_t *cls)
{
    category_list *cats;
    BOOL isMeta;

    rwlock_assert_writing(&runtimeLock);

    isMeta = isMetaClass(cls);

    // Re-methodizing: check for more categories
    if ((cats = unattachedCategoriesForClass(cls))) {
        chained_property_list *newproperties;
        const protocol_list_t **newprotos;

        if (PrintConnecting) {
            _objc_inform("CLASS: attaching categories to class '%s' %s",
                         getName(cls), isMeta ? "(meta)" : "");
        }

        // Update methods, properties, protocols

        BOOL vtableAffected = NO;
        attachCategoryMethods(cls, cats, &vtableAffected);

        newproperties = buildPropertyList(NULL, cats, isMeta);
        if (newproperties) {
            newproperties->next = cls->data()->properties;
            cls->data()->properties = newproperties;
        }

        newprotos = buildProtocolList(cats, NULL, cls->data()->protocols);
        if (cls->data()->protocols  &&  cls->data()->protocols != newprotos) {
            _free_internal(cls->data()->protocols);
        }
        cls->data()->protocols = newprotos;

        _free_internal(cats);

        // Update method caches and vtables
        flushCaches(cls);
        if (vtableAffected) flushVtables(cls);
    }
}

The function integrates methods, attributes and protocols in category into main class/metaclass, updates methods_lists, properties and protocols in data field (), and calls attachCategory Methods for adding instance methods.

static void 
attachCategoryMethods(class_t *cls, category_list *cats,
                      BOOL *inoutVtablesAffected)
{
    if (!cats) return;
    if (PrintReplacedMethods) printReplacements(cls, cats);

    BOOL isMeta = isMetaClass(cls);
    method_list_t **mlists = (method_list_t **)
        _malloc_internal(cats->count * sizeof(*mlists));

    // Count backwards through cats to get newest categories first
    int mcount = 0;
    int i = cats->count;
    BOOL fromBundle = NO;
    while (i--) {
        method_list_t *mlist = cat_method_list(cats->list[i].cat, isMeta);
        if (mlist) {
            mlists[mcount++] = mlist;
            fromBundle |= cats->list[i].fromBundle;
        }
    }

    attachMethodLists(cls, mlists, mcount, NO, fromBundle, inoutVtablesAffected);

    _free_internal(mlists);

}

Its work can be seen as putting together the list of instance methods of category into a list of strength methods, and then calling attachMethodLists for processing.

for (uint32_t m = 0;
             (scanForCustomRR || scanForCustomAWZ)  &&  m < mlist->count;
             m++)
        {
            SEL sel = method_list_nth(mlist, m)->name;
            if (scanForCustomRR  &&  isRRSelector(sel)) {
                cls->setHasCustomRR();
                scanForCustomRR = false;
            } else if (scanForCustomAWZ  &&  isAWZSelector(sel)) {
                cls->setHasCustomAWZ();
                scanForCustomAWZ = false;
            }
        }

        // Fill method list array
        newLists[newCount++] = mlist;
    .
    .
    .

    // Copy old methods to the method list array
    for (i = 0; i < oldCount; i++) {
        newLists[newCount++] = oldLists[i];
    }

Here we need to note that category does not completely replace the same name method of the original class. The category method is placed in front of the new method list, while the original class method is placed behind. In runtime, when the method of category is found, the method of category will stop traversing. This is what we usually call "covering" method.

Finding the source method is very simple, just traverse the list of methods and find the last corresponding name method.

Class currentClass = [MyClass class];
MyClass *my = [[MyClass alloc] init];

if (currentClass) {
    unsigned int methodCount;
    Method *methodList = class_copyMethodList(currentClass, &methodCount);
    IMP lastImp = NULL;
    SEL lastSel = NULL;
    for (NSInteger i = 0; i < methodCount; i++) {
        Method method = methodList[i];
        NSString *methodName = [NSString stringWithCString:sel_getName(method_getName(method)) 
                                        encoding:NSUTF8StringEncoding];
        if ([@"printName" isEqualToString:methodName]) {
            lastImp = method_getImplementation(method);
            lastSel = method_getName(method);
        }
    }
    typedef void (*fn)(id,SEL);

    if (lastImp != NULL) {
        fn f = (fn)lastImp;
        f(my,lastSel);
    }
    free(methodList);
}

category object association

So when using object associations in categorization, what is the corresponding storage location and lifecycle?
To explore the source code, in the objc-references.mm file void_object_set_associative_reference method

void _object_set_associative_reference(id object, void *key, id value, uintptr_t policy) {
    // retain the new value (if any) outside the lock.
    ObjcAssociation old_association(0, nil);
    id new_value = value ? acquireValue(value, policy) : nil;
    {
        AssociationsManager manager;
        AssociationsHashMap &associations(manager.associations());
        disguised_ptr_t disguised_object = DISGUISE(object);
        if (new_value) {
            // break any existing association.
            AssociationsHashMap::iterator i = associations.find(disguised_object);
            if (i != associations.end()) {
                // secondary table exists
                ObjectAssociationMap *refs = i->second;
                ObjectAssociationMap::iterator j = refs->find(key);
                if (j != refs->end()) {
                    old_association = j->second;
                    j->second = ObjcAssociation(policy, new_value);
                } else {
                    (*refs)[key] = ObjcAssociation(policy, new_value);
                }
            } else {
                // create the new association (first time).
                ObjectAssociationMap *refs = new ObjectAssociationMap;
                associations[disguised_object] = refs;
                (*refs)[key] = ObjcAssociation(policy, new_value);
                object->setHasAssociatedObjects();
            }
        } else {
            // setting the association to nil breaks the association.
            AssociationsHashMap::iterator i = associations.find(disguised_object);
            if (i !=  associations.end()) {
                ObjectAssociationMap *refs = i->second;
                ObjectAssociationMap::iterator j = refs->find(key);
                if (j != refs->end()) {
                    old_association = j->second;
                    refs->erase(j);
                }
            }
        }
    }
    // release the old value (outside of the lock).
    if (old_association.hasValue()) ReleaseValue()(old_association);
}

We can see that the associated object is managed by Associations Manager, so what is it?

/ class AssociationsManager manages a lock / hash table singleton pair.
// Allocating an instance acquires the lock, and calling its assocations()
// method lazily allocates the hash table.

spinlock_t AssociationsManagerLock;

class AssociationsManager {
    // associative references: object pointer -> PtrPtrHashMap.
    static AssociationsHashMap *_map;
public:
    AssociationsManager()   { AssociationsManagerLock.lock(); }
    ~AssociationsManager()  { AssociationsManagerLock.unlock(); }
    
    AssociationsHashMap &associations() {
        if (_map == NULL)
            _map = new AssociationsHashMap();
        return *_map;
    }
};

AssociationsHashMap *AssociationsManager::_map = NULL;

Associations Manager is a static global associations HashMap that stores all associated objects, key is the memory address of the object, value is another associations HashMap that stores the kv of the associated object, and object destruction is handed over to objc_destruct Instance.

void *objc_destructInstance(id obj) 
{
    if (obj) {
        Class isa_gen = _object_getClass(obj);
        class_t *isa = newcls(isa_gen);

        // Read all of the flags at once for performance.
        bool cxx = hasCxxStructors(isa);
        bool assoc = !UseGC && _class_instancesHaveAssociatedObjects(isa_gen);

        // This order is important.
        if (cxx) object_cxxDestruct(obj);
        if (assoc) _object_remove_assocations(obj);

        if (!UseGC) objc_clear_deallocating(obj);
    }

    return obj;
}

If there are any mistakes, please correct them.~

Reference link:
https://tech.meituan.com/DiveIntoCategory.html
http://blog.leichunfeng.com/blog/2015/05/18/objective-c-category-implementation-principle/
https://developer.apple.com/library/ios/documentation/General/Conceptual/DevPedia-CocoaCore/Category.html#//apple_ref/doc/uid/TP40008195-CH5-SW1 http://stackoverflow.com/questions/5272451/overriding-methods-using-categories-in-objective-c

Here you can see the source code