转自公众号:NA分享
performSelector简单使用
iOS中提供了如下几种常用的调用方式:
[self performSelector:@selector(sureTestMethod)];
[self performSelector:@selector(sureTestMethod)
withObject:params];
[self performSelector:@selector(sureTestMethod)
withObject:params
withObject:params2];
iOS动态调用
performSelector
可以向一个对象传递任何消息,而不需要在编译的时候声明这些方法,这也是runtime
的一种应用方式,所以performSelector
和直接调用方法的区别就在与runtime
。直接调用编译是会自动校验。如果方法不存在,那么直接调用 在编译时候就能够发现,编译器会直接报错。 但是使用performSelector
的话一定是在运行时候才能发现,如果此方法不存在就会崩溃。所以performSelector
和- (BOOL)respondsToSelector:(SEL)aSelector
搭配使用,来在运行时判断对象是否响应此方法。因为此特性,performSelector
也广泛用于动态化和组件化的模块中。
SEL selector = @selector(dynamicMethod);
[self performSelector:selector];
如果方法名称也是动态不确定的,会提示如下警告:
⚠️ PerformSelector may cause a leak because its selector is unknown
意为因为当前方法名未知可能会引起内存泄露相关问题。 可以通过如下代码忽略此警告
#pragma clang diagnostic push
#pragma clang diagnostic ignored "-Warc-performSelector-leaks"
[self performSelector:selector];
#pragma clang diagnostic pop
performSelector
默认最多只可传递两个参数,若需多参参考以下方式:
一种是使用NSInvocation
,利用了runtime
的反射机制,效率较低,可读性不高;
第二种是将参数封装进NSArray
、NSDictionary
等对象,可读性强,效率高;
第三种是使用objc_msgSend
重写performSelector
。
第一种:NSInvocation
- (id)performSelector:(SEL)selector withObjects:(NSArray *)objects
{
// 方法签名(方法的描述)
NSMethodSignature *signature = [[self class] instanceMethodSignatureForSelector:selector];
if (signature == nil) {
//可以抛出异常也可以不操作。
}
// NSInvocation : 利用一个NSInvocation对象包装一次方法调用(方法调用者、方法名、方法参数、方法返回值)
NSInvocation *invocation = [NSInvocation invocationWithMethodSignature:signature];
invocation.target = self;
invocation.selector = selector;
// 设置参数
NSInteger paramsCount = signature.numberOfArguments - 2; // 除self、_cmd以外的参数个数
paramsCount = MIN(paramsCount, objects.count);
for (NSInteger i = 0; i < paramsCount; i++) {
id object = objects[i];
if ([object isKindOfClass:[NSNull class]]) continue;
[invocation setArgument:&object atIndex:i + 2];
}
// 调用方法
[invocation invoke];
// 获取返回值
id returnValue = nil;
if (signature.methodReturnLength) { // 有返回值类型,才去获得返回值
[invocation getReturnValue:&returnValue];
}
return returnValue;
}
- (void)test {
NSString *str = @"字符串";
NSNumber *num = @20;
NSArray *arr = @[@"数组值1", @"数组值2"];
SEL sel = NSSelectorFromString(@"NSInvocationWithString:withNum:withArray:");
NSArray *objs = [NSArray arrayWithObjects:str, num, arr, nil];
[self performSelector:sel withObjects:objs];
}
- (void)NSInvocationWithString:(NSString *)string withNum:(NSNumber *)number withArray:(NSArray *)array {
NSLog(@"%@, %@, %@", string, number, array[0]);
}
第二种:省略...
第三种:objc_msgSend
- (void)test {
NSString *str = @"字符串objc_msgSend";
NSNumber *num = @20;
NSArray *arr = @[@"数组值1", @"数组值2"];
SEL sel = NSSelectorFromString(@"ObjcMsgSendWithString:withNum:withArray:");
((void (*) (id, SEL, NSString *, NSNumber *, NSArray *)) objc_msgSend) (self, sel, str, num, arr);
}
- (void)ObjcMsgSendWithString:(NSString *)string withNum:(NSNumber *)number withArray:(NSArray *)array {
NSLog(@"%@, %@, %@", string, number, array[0]);
}
如果参数中有结构体怎么办?
可以把结构体转换成对象。
typedef struct ParameterStruct{
int a;
int b;
}MyStruct;
- (void)test {
NSString *str = @"字符串 把结构体转换为对象";
NSNumber *num = @20;
NSArray *arr = @[@"数组值1", @"数组值2"];
MyStruct mystruct = {10,20};
NSValue *value = [NSValue valueWithBytes:&mystruct objCType:@encode(MyStruct)];
SEL sel = NSSelectorFromString(@"NSInvocationWithString:withNum:withArray:withValue:");
NSArray *objs = [NSArray arrayWithObjects:str, num, arr, value,nil];
[self performSelector:sel withObjects:objs]; // 同第一种方法
}
- (void)NSInvocationWithString:(NSString *)string withNum:(NSNumber *)number withArray:(NSArray *)array withValue:(NSValue *)value{
MyStruct struceBack;
[value getValue:&struceBack];
NSLog(@"%@, %@, %@, %d", string, number, array[0],struceBack.a);
}
转自公众号:NA分享