1.小文件的下载
由于文件较小,我们可以直接可以使用NSURLConnection的异步请求(默认在这里开了一条线程,不回阻塞主线程)方法,在block里面进行数据接收:
[NSURLConnection sendAsynchronousRequest:request
queue:[NSOperationQueue mainQueue]
completionHandler:^(NSURLResponse * _Nullable response, NSData * _Nullable data, NSError * _Nullable connectionError) {
if (!connectionError) {
dispatch_async(dispatch_get_main_queue(), ^{
_imageView.image = [[UIImage alloc] initWithData:data];
});
} else {
NSLog(@"erorr = %@",[connectionError description]);
}
}];
可是大文件下载我们一般就不采用这种方式了,这种方式下载文件看不到进度,并且虽然在这里开了子线程来进行下载,可是在主线程的UI刷新是要等到服务器返回了总的文件data才会进行,这就给用户体验曹成了不好的地方,而且在大的文件下载过程中有可能会涉及到断点下载的情况(比如断网或者手动暂停)。
2.大文件的下载
具体思路是初始化一个NSMutableData类型的_responseData,在下载的过程中在- (void)connection:(NSURLConnection *)connection didReceiveData:(NSData *)data
代理方法里面进行数据拼接,每次接收到数据就append,在代理方法- (void)connection:(NSURLConnection *)connection didReceiveResponse:(NSURLResponse *)response {
里面进行文件的写入,至于进度监听,可以在每次接收到数据的时候算目前接受到的总长度_currentDownLoadLength,然后除以文件的response.expectedContentLength,断点下载的关键是设置当前请求的请求头:
//设置请求头
NSString *range = [NSString stringWithFormat:@"bytes=%lld-",_currentDownLoadLength];
[request setValue:range forHTTPHeaderField:@"Range"];
这里有一个新的东西:http协议请求头里面的Range,Range是请求头里面的下载起点(或者说下载的进度范围 ):
//Range 可以指定每次从网络下载数据包的大小
bytes = 0 - 499 //从0到499共500
bytes = 500 - //从500到结束
bytes = -500 //最后500
bytes = 500 - 599, 800 - 899 //同时指定几个范围
我们可以在暂停或者断网的情况下,记录当前的下载总长度,然后设置新的请求头,并重新初始化一个NSURLConnection对象,并在代理方法里面继续进行数据接收。
具体实现代理如下:
#pragma mark --NSURLConnectionDataDelegate
- (void)connectionDidFinishLoading:(NSURLConnection *)connection {
NSString *dbPath = [[NSSearchPathForDirectoriesInDomains(NSDocumentDirectory, NSUserDomainMask, YES) firstObject] stringByAppendingPathComponent:_fileName];
//下载完成 写入文件
[_responseData writeToFile:dbPath atomically:YES];
}
- (void)connection:(NSURLConnection *)connection didReceiveResponse:(NSURLResponse *)response {
_fileName = response.suggestedFilename;
_totalLength = response.expectedContentLength;
NSLog(@"fileName = %@",response.suggestedFilename);
}
- (void)connection:(NSURLConnection *)connection didReceiveData:(NSData *)data {
//拼接每次接受的数据
[_responseData appendData:data];
//计算目前接收到的数据总长度
_currentDownLoadLength = _currentDownLoadLength + data.length;
//下载进度百分比
NSString *subStr = @"%";
NSLog(@"进度 = %@%.2f",subStr, 100.0 *_currentDownLoadLength /_totalLength);
}
#pragma mark --private Method
- (void)buttonAction:(id)sender {
UIButton *button = (UIButton *)sender;
button.selected = !button.selected;
//断点下载
if (button.selected) {
//设置请求头
NSString *range = [NSString stringWithFormat:@"bytes=%lld-",_currentDownLoadLength];
NSMutableURLRequest *request = [NSMutableURLRequest requestWithURL:[NSURL URLWithString:@"http://download.xmcdn.com/group18/M01/BC/91/wKgJKlfAEN6wZgwhANQvLrUQ3Pg146.aac"]];
[request setValue:range forHTTPHeaderField:@"Range"];
//重新请求
_connection = [[NSURLConnection alloc] initWithRequest:request
delegate:self];
[_connection start];
} else {
//暂停
[_connection cancel];
_connection = nil;
}
}
下载完成之后将_repsonseData写入sandBox,效果如下:
NSURLConnection下载.png
那么问题又来了,有些童鞋可能会发现,加入一个文件很大的情况下,会出现内存暴涨的情况,从而使程序报memory warning的情况,这样的代码是不完整的,也是质量不高的,于是我们要解决内存暴涨的问题
3.解决大文件内存暴涨问题
思路:我们注意到_responseData存在于内存中,每次append的是时候,内存就逐渐暴涨,于是我们可以开多个子线程来下载同一个文件,并且在下载的过程中可以边下载边写入沙盒sandBox,这就需要在接收到服务器响应的时候,开多个connection对象来下载(平分区域设置请求头的下载范围),使用NSFileHandel文件句柄写入沙盒,在finishload代理方法里面进行文件的合并处理,具体实现代码如下:
#pragma mark - NSURLConnectionDataDelegate
- (void)connection:(NSURLConnection *)connection didReceiveResponse:(NSURLResponse *)response {
if ([connection isEqual:_connection]) {
//获取总长度
_totalWriteDataLength = response.expectedContentLength;
//取消
[_connection cancel];
NSString *cachePath = NSSearchPathForDirectoriesInDomains(NSCachesDirectory, NSUserDomainMask, YES).lastObject;
//开4个connection同时进行下载
for (int i = 0; i < 4; i++) {
//4个文件
NSString *filePath = [cachePath stringByAppendingPathComponent:[NSString stringWithFormat:@"%@_%d", response.suggestedFilename, i]];
//创建4个临时文件
NSFileManager *fileManager = [NSFileManager defaultManager];
[fileManager createFileAtPath:filePath contents:nil attributes:nil];
//创建请求
NSMutableURLRequest *request = [NSMutableURLRequest requestWithURL:[NSURL URLWithString:@"http://download.xmcdn.com/group18/M01/BC/91/wKgJKlfAEN6wZgwhANQvLrUQ3Pg146.aac"]];
//平分每个Connection的下载范围
NSString *range = [NSString stringWithFormat:@"bytes=%lld-%lld", response.expectedContentLength/4*i, response.expectedContentLength/4*(i+1)];
[request setValue:range forHTTPHeaderField:@"Range"];
//创建多个请求
NSURLConnection *connection = [NSURLConnection connectionWithRequest:request delegate:self];
//创建4个文件句柄
NSFileHandle *fileHandle = [NSFileHandle fileHandleForWritingAtPath:filePath];
[_pathArray addObject:filePath];
[_connectionArray addObject:connection];
[_fileHandleArry addObject:fileHandle];
}
}
}
在finishload代理方法里面进行文件合并处理:
- (void)connectionDidFinishLoading:(NSURLConnection *)connection {
_finishedCount++;
NSInteger index = [_connectionArray indexOfObject:connection];
//获取句柄
NSFileHandle *fileHandle = [_fileHandleArry objectAtIndex:index];
[fileHandle closeFile];
fileHandle = nil;
if (_finishedCount == 4)//将4个任务下载的文件合并成一个
{
NSFileManager *fileManager = [NSFileManager defaultManager];
NSString *tmpPath = [_pathArray objectAtIndex:index];
NSString *filePath = [tmpPath substringToIndex:tmpPath.length];
[fileManager createDirectoryAtPath:filePath withIntermediateDirectories:YES attributes:nil error:nil];
//创建一个文件句柄
NSFileHandle *fileHandle = [NSFileHandle fileHandleForWritingAtPath:filePath];
for (int i = 0; i < 4; i++) {
[fileHandle seekToEndOfFile];
//向总文件写入数据
[fileHandle writeData:[NSData dataWithContentsOfFile:[_pathArray objectAtIndex:i]]];
}
[fileHandle closeFile];
fileHandle = nil;
}
}
我们可以看内存不会存在暴涨的情况了,说明了这一些代码是有意义的,目前关于NSURLConnection的下载介绍就是这些,由于本人技术有限,写的不好的地方请大家指正,如果觉得本人对于您有帮助的话,请动用您宝贵的双手点个赞,谢谢!
作者------mrChan1234
网友评论