weakSelf+strongSelf

本文深入探讨了Objective-C中弱引用(__weak)与强引用(__strong)的使用场景与作用。通过具体示例解释了在block外使用__weak typeof(self) weakSelf = self;的作用是对象释放时自动对weak指针置nil,防止循环引用。而在block内使用__strong typeof(weakSelf) strongSelf = weakSelf;则保证了self在block执行过程中始终存在。

摘要生成于 C知道 ,由 DeepSeek-R1 满血版支持, 前往体验 >

 block外:__weak __typeof(self)weakSelf = self;

作用:对象释放时,自动对weak指针置nil,self不知道什么时候释放;

block内:__strong __typeof(weakSelf) strongSelf = weakSelf;

作用:保证self在block执行过程中,一直存在;

 

 

处理下面代码可能出现线程问题的情况- (void)shutterCamera { AVCaptureConnection *videoConnection = [self.ImageOutPut connectionWithMediaType:AVMediaTypeVideo]; UIDeviceOrientation curDeviceOrientation = [[UIDevice currentDevice] orientation]; if (!videoConnection || !self.session.isRunning) { ATLog(@"take photo failed!"); return; } AVCaptureVideoOrientation avcaptureOrientation = [self avOrientationForDeviceOrientation:curDeviceOrientation]; [videoConnection setVideoOrientation:avcaptureOrientation]; [videoConnection setVideoScaleAndCropFactor:self.effectiveScale]; MJWeakSelf; [self.ImageOutPut captureStillImageAsynchronouslyFromConnection:videoConnection completionHandler:^(CMSampleBufferRef imageDataSampleBuffer, NSError *error) { dispatch_sync(dispatch_get_global_queue(DISPATCH_QUEUE_PRIORITY_DEFAULT, 0), ^{ __strong typeof(weakSelf) strongSelf = weakSelf; if (!strongSelf) return; if (imageDataSampleBuffer) { NSData * imageData = [AVCaptureStillImageOutput jpegStillImageNSDataRepresentation:imageDataSampleBuffer]; strongSelf.image = [UIImage imageWithData:imageData]; if (strongSelf.image) { strongSelf.image = [UIImage updateImageSize:strongSelf.image]; } strongSelf.image = [[TZImageManager manager] fixOrientation:strongSelf.image];//[UIImage fixOrientation:weakSelf.image]; PHAuthorizationStatus authorStatus = [PHPhotoLibrary authorizationStatus]; if (authorStatus == PHAuthorizationStatusAuthorized) { [[TZImageManager manager] savePhotoWithImage:strongSelf.image completion:^(PHAsset *asset, NSError *error) { [strongSelf addPHAsset:asset]; }]; }else{ [strongSelf gotoEditImageController:strongSelf.image asset:nil]; } } }); }]; } - (void)gotoEditImageController:(UIImage *)image asset:(PHAsset *)asset { if (!image) { [ATMBPHud showErrorToView:self.view text:@"获取图片失败,请重新再试"]; return; } if (image) { image = [UIImage updateImageSize:image]; } YasicClipPage *vc = [[YasicClipPage alloc] init]; vc.userId = self.userId; vc.attention = self.attention; vc.targetImage = image; vc.searchType = self.searchType; vc.asset = asset; // [self stopSession]; dispatch_async(dispatch_get_main_queue(), ^{ [[KDUserInterfaceTool currenrtopViewController].navigationController pushViewController:vc animated:YES]; [self removeViewController:self]; }); } - (void)addPHAsset:(PHAsset *)asset { MJWeakSelf; TZAssetModel *assetModel = [[TZImageManager manager] createModelWithAsset:asset]; [ATHUD show]; [[TZImageManager manager] requestImageDataForAsset:assetModel.asset completion:^(NSData *imageData, NSString *dataUTI, UIImageOrientation orientation, NSDictionary *info) { [ATHUD dismiss]; UIImage *image = [UIImage imageWithData:imageData];//[[TZImageManager manager] fixOrientation:[UIImage imageWithData:imageData]]; image = [[TZImageManager manager] fixOrientation:image]; [weakSelf gotoEditImageController:image asset:asset]; } progressHandler:^(double progress, NSError *error, BOOL *stop, NSDictionary *info) { ATLog(@"%f",progress); }]; } #pragma --mark 获取大图 - (void)getImageWithModel:(TZAssetModel *)model { if (model == nil) { ATLog(@"图片为空"); return; } [ATHUD show]; WS(weakSelf) [[TZImageManager manager] requestImageDataForAsset:model.asset completion:^(NSData *imageData, NSString *dataUTI, UIImageOrientation orientation, NSDictionary *info) { [ATHUD dismiss]; UIImage *image = [UIImage imageWithData:imageData]; if (weakSelf.downBtnClick) { weakSelf.downBtnClick(image, model); } } progressHandler:^(double progress, NSError *error, BOOL *stop, NSDictionary *info) { ATLog(@"%f",progress); }]; }
07-25
内容概要:《中文大模型基准测评2025年上半年报告》由SuperCLUE团队发布,详细评估了2025年上半年中文大模型的发展状况。报告涵盖了大模型的关键进展、国内外大模型全景图及差距、专项测评基准介绍等。通过SuperCLUE基准,对45个国内外代表性大模型进行了六大任务(数学推理、科学推理、代码生成、智能体Agent、精确指令遵循、幻觉控制)的综合测评。结果显示,海外模型如o3、o4-mini(high)在推理任务上表现突出,而国内模型如Doubao-Seed-1.6-thinking-250715在智能体Agent和幻觉控制任务上表现出色。此外,报告还分析了模型性价比、效能区间分布,并对代表性模型如Doubao-Seed-1.6-thinking-250715、DeepSeek-R1-0528、GLM-4.5等进行了详细介绍。整体来看,国内大模型在特定任务上已接近国际顶尖水平,但在综合推理能力上仍有提升空间。 适用人群:对大模型技术感兴趣的科研人员、工程师、产品经理及投资者。 使用场景及目标:①了解2025年上半年中文大模型的发展现状与趋势;②评估国内外大模型在不同任务上的表现差异;③为技术选型和性能优化提供参考依据。 其他说明:报告提供了详细的测评方法、评分标准及结果分析,确保评估的科学性和公正性。此外,SuperCLUE团队还发布了多个专项测评基准,涵盖多模态、文本、推理等多个领域,为业界提供全面的测评服务。
评论
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值