Assuming you're looking for a quick tactical fix, what you need to do is make sure the cell image is initialized and also that the cell's row is still visible, e.g:
- (UITableViewCell *)tableView:(UITableView *)tableView cellForRowAtIndexPath:(NSIndexPath *)indexPath {
MyCell *cell = [tableView dequeueReusableCellWithIdentifier:@"cell" forIndexPath:indexPath];
cell.poster.image = nil; // or cell.poster.image = [UIImage imageNamed:@"placeholder.png"];
NSURL *url = [NSURL URLWithString:[NSString stringWithFormat:@"http://myurl.com/%@.jpg", self.myJson[indexPath.row][@"movieId"]]];
NSURLSessionTask *task = [[NSURLSession sharedSession] dataTaskWithURL:url completionHandler:^(NSData * _Nullable data, NSURLResponse * _Nullable response, NSError * _Nullable error) {
if (data) {
UIImage *image = [UIImage imageWithData:data];
if (image) {
dispatch_async(dispatch_get_main_queue(), ^{
MyCell *updateCell = (id)[tableView cellForRowAtIndexPath:indexPath];
if (updateCell)
updateCell.poster.image = image;
});
}
}
}];
[task resume];
return cell;
}
The above code addresses a few problems stemming from the fact that the cell is reused:
You're not initializing the cell image before initiating the background request (meaning that the last image for the dequeued cell will still be visible while the new image is downloading). Make sure to nil
the image
property of any image views or else you'll see the flickering of images.
A more subtle issue is that on a really slow network, your asynchronous request might not finish before the cell scrolls off the screen. You can use the UITableView
method cellForRowAtIndexPath:
(not to be confused with the similarly named UITableViewDataSource
method tableView:cellForRowAtIndexPath:
) to see if the cell for that row is still visible. This method will return nil
if the cell is not visible.
The issue is that the cell has scrolled off by the time your async method has completed, and, worse, the cell has been reused for another row of the table. By checking to see if the row is still visible, you'll ensure that you don't accidentally update the image with the image for a row that has since scrolled off the screen.
Somewhat unrelated to the question at hand, I still felt compelled to update this to leverage modern conventions and API, notably:
Use NSURLSession
rather than dispatching -[NSData contentsOfURL:]
to a background queue;
Use dequeueReusableCellWithIdentifier:forIndexPath:
rather than dequeueReusableCellWithIdentifier:
(but make sure to use cell prototype or register class or NIB for that identifier); and
I used a class name that conforms to Cocoa naming conventions (i.e. start with the uppercase letter).
Even with these corrections, there are issues:
The above code is not caching the downloaded images. That means that if you scroll an image off screen and back on screen, the app may try to retrieve the image again. Perhaps you'll be lucky enough that your server response headers will permit the fairly transparent caching offered by NSURLSession
and NSURLCache
, but if not, you'll be making unnecessary server requests and offering a much slower UX.
We're not canceling requests for cells that scroll off screen. Thus, if you rapidly scroll to the 100th row, the image for that row could be backlogged behind requests for the previous 99 rows that aren't even visible anymore. You always want to make sure you prioritize requests for visible cells for the best UX.
The simplest fix that addresses these issues is to use a UIImageView
category, such as is provided with SDWebImage or AFNetworking. If you want, you can write your own code to deal with the above issues, but it's a lot of work, and the above UIImageView
categories have already done this for you.
与恶龙缠斗过久,自身亦成为恶龙;凝视深渊过久,深渊将回以凝视…