Re: Memory Warnings
(reposted after editing for length)
On Oct 7, 2014, at 22:12 , Gerriet M. Denkmann gerr...@mdenkmann.de wrote:
Is there a way for the app to find out how much memory it is currently using?
Here’s my take on this, which may be way off base:
There’s really no practical way to answer such questions any more. The way
memory is managed is so varied, so diffused that even if you could get an
answer, it would be as likely to lead you astray as to help you.
In practice, I think the question is not how much memory your app is using, but
how fast your memory use is increasing over time.
If it’s not increasing over time — and I’m speaking in the most general way
possible here, so using more memory per opened document isn’t what I’d call
increasing over time, in this context — then you have a fairly easy evaluation.
Your app fits (in the foreground) or it crashes. Your app retreats gracefully
(to the background) or it’s terminated.
The way we program now, we generally regard increasing memory use over time as
a memory leak, and we fix our code so this doesn’t happen and usage is stable.
So, it fits or it crashes.
When I have a file of 100 KB and do:
myData = [NSData dataWithContentsOfFile:options:error: ];
myBytes = myData.bytes;
how much Ram am I now using? 100 KB or 200 KB? I.e. does bytes just return
a pointer to some internal structure in NSData?
Would options = NSDataReadingMappedIfSafe be of some help?
The general answer is that you don’t know how much it’s using. In practice, the
answer used to be 100KB (“bytes” returned an interior pointer). But this isn’t
always true any more. GCD data blocks, which can be bridged to NSData, can come
in pieces, and in that case “bytes” may copy the data into contiguous memory.
Mapped data might help in common cases, but how does this save you from “it
fits or it crashes” when the file turns out to be on a non-local file system,
on which mapping doesn’t work? Memory mapping is more about performance than
size, perhaps.
static dispatch_once_t justOnce;
dispatch_once( justOnce, ^{ _myData = [NSData
dataWithContentsOfFile:options:error: ]; } );
If I would (upon receiving a memory warning) release myData, I would never be
able to get it back, would I?
How could I handle this case?
Given that ‘dispatch_once’ is intended to do something only once, it’s hardly a
shock that it doesn’t help in this case. Anyway, the problem is not in keeping
a single but discardable copy of the data, it’s in making the discard/recovery
process thread safe.
You can easily use a construct like GCD semaphores to ensure the integrity of
the data recovery after discarding, but you have to be careful to avoid
deadlocks. Or, you can easily use a construct like GCD serial queues (perhaps
even the main queue) as a deadlock-free of getting to the same result, but you
might have to redesign your code for asynchronicity with regard to the queuing.
Once you’re in the deeps of thread safety, I’m pretty sure there are no more
free lunches.
___
Cocoa-dev mailing list (Cocoa-dev@lists.apple.com)
Please do not post admin requests or moderator comments to the list.
Contact the moderators at cocoa-dev-admins(at)lists.apple.com
Help/Unsubscribe/Update your Subscription:
https://lists.apple.com/mailman/options/cocoa-dev/archive%40mail-archive.com
This email sent to arch...@mail-archive.com
Re: Memory Warnings
On 8 Oct 2014, at 13:42, Quincey Morris quinceymor...@rivergatesoftware.com wrote: (reposted after editing for length) On Oct 7, 2014, at 22:12 , Gerriet M. Denkmann gerr...@mdenkmann.de wrote: Is there a way for the app to find out how much memory it is currently using? Here’s my take on this, which may be way off base: There’s really no practical way to answer such questions any more. The way memory is managed is so varied, so diffused that even if you could get an answer, it would be as likely to lead you astray as to help you. In practice, I think the question is not how much memory your app is using, but how fast your memory use is increasing over time. If it’s not increasing over time It's not. Checked also with Leaks Instrument. — then you have a fairly easy evaluation. Your app fits (in the foreground) or it crashes. Your app retreats gracefully (to the background) or it’s terminated. It runs in the foreground. But when in background it sometimes gets terminated. When I have a file of 100 KB and do: myData = [NSData dataWithContentsOfFile:options:error: ]; myBytes = myData.bytes; how much Ram am I now using? 100 KB or 200 KB? I.e. does bytes just return a pointer to some internal structure in NSData? Would options = NSDataReadingMappedIfSafe be of some help? The general answer is that you don’t know how much it’s using. In practice, the answer used to be 100KB (“bytes” returned an interior pointer). But this isn’t always true any more. GCD data blocks, which can be bridged to NSData, can come in pieces, and in that case “bytes” may copy the data into contiguous memory. Mapped data might help in common cases, but how does this save you from “it fits or it crashes” when the file turns out to be on a non-local file system, on which mapping doesn’t work? Memory mapping is more about performance than size, perhaps. I made a small iOS app which has just 3 buttons: Data (loads NSData from file, which resides in the app bundle) Bytes (which uses myData.bytes) Count (which sums all bytes in myData) Result with NSDataReadingMappedIfSafe, Persistent Bytes in Instruments, Allocations: No change with any button (always ca. 3.5 MB). Data button takes 1 msec, Bytes 73 μsec, first Count takes 250 msec longer than subsequent Counts. Without NSDataReadingMappedIfSafe: Data button increases Ram use by 15 MB (about size of file); takes 250 msec. Bytes, Count: no change. Bytes takes 13 μsec. All Counts take same time (ca. 50 msec). So NSDataReadingMappedIfSafe seems to be a good idea (at least in my case of a fixed file in the app bundle). You can easily use a construct like GCD semaphores... Sounds too much work for now. Maybe to be considered later. Thanks for your ideas though. Kind regards, Gerriet. ___ Cocoa-dev mailing list (Cocoa-dev@lists.apple.com) Please do not post admin requests or moderator comments to the list. Contact the moderators at cocoa-dev-admins(at)lists.apple.com Help/Unsubscribe/Update your Subscription: https://lists.apple.com/mailman/options/cocoa-dev/archive%40mail-archive.com This email sent to arch...@mail-archive.com
Re: Responsive scrolling: prepareContentInRect: not called
On 08 Oct 2014, at 01:41, Ken Thomases k...@codeweavers.com wrote: Wondering if anyone has an idea of why prepareContentInRect: is not called for a view when responsive scrolling is enabled. We are using an NSScrollView subclass that overrides prepareContentInRect:. I would not expect -prepareContentInRect: to be called on the scroll view. It would be called on the document view (and its subviews?). The purpose is to allow the document view to prepare itself for the requests to draw the overdraw region. The scroll view itself is not drawn in the overdraw region. Likewise, I would not expect the scroll view's -drawRect: to be called for overdrawing. Thanks for your suggestions Ken, Indeed we do not ‘drawRect:’ anything in the scroll view. :-) (I have tried overriding prepareContentInRect: in the NSScrollView’s subclass, as well as in the view that serves as the scroll view’s document view. Neither of these methods gets called though.) I would double-check that, in the occasions when you overrode -prepareContentInRect: in the document view, you didn't also accidentally disable responsive scrolling. Also, make sure your override calls through to super and doesn't unintentionally stop AppKit from further extending the overdraw region by passing the same rect in successive calls. Turns out the problem is a different, and of course, almost embarrassingly silly one. I misspelled the method name: prepareContentInrect rather than prepareContentInRect. Now that I’ve caught that and corrected it, the method is called. Thanks for you effort anyway. -António ___ Cocoa-dev mailing list (Cocoa-dev@lists.apple.com) Please do not post admin requests or moderator comments to the list. Contact the moderators at cocoa-dev-admins(at)lists.apple.com Help/Unsubscribe/Update your Subscription: https://lists.apple.com/mailman/options/cocoa-dev/archive%40mail-archive.com This email sent to arch...@mail-archive.com
set maximum file size
Hi folks I have an app which needs to save some of its data into a log file. I have the method set up already for creating and appending the log file. However, I'd like to limit the file size and initiate a file turn over when it reaches a certain size. What is the best way to accomplish this? I'm aware of Cocoa Lumberjack, but I really don't want to import an entire 3rd party framework for the sake of the one simple function that I need if I can avoid it. Any thoughts much appreciated. Best P ___ Cocoa-dev mailing list (Cocoa-dev@lists.apple.com) Please do not post admin requests or moderator comments to the list. Contact the moderators at cocoa-dev-admins(at)lists.apple.com Help/Unsubscribe/Update your Subscription: https://lists.apple.com/mailman/options/cocoa-dev/archive%40mail-archive.com This email sent to arch...@mail-archive.com
Re: set maximum file size
Why not check the file size before writing? If the file size your max, then get the length of the message you are about to write and trim that many chars from the front of the file before writing. Or append the message to your file and after writing, check the file size and trim the surplus chats from the beginning. Do those suggestions work for you? Sent from my iPod Touch. Please pardon typos. On Oct 8, 2014, at 10:01 AM, 2551 2551p...@gmail.com wrote: Hi folks I have an app which needs to save some of its data into a log file. I have the method set up already for creating and appending the log file. However, I'd like to limit the file size and initiate a file turn over when it reaches a certain size. What is the best way to accomplish this? I'm aware of Cocoa Lumberjack, but I really don't want to import an entire 3rd party framework for the sake of the one simple function that I need if I can avoid it. Any thoughts much appreciated. Best P ___ Cocoa-dev mailing list (Cocoa-dev@lists.apple.com) Please do not post admin requests or moderator comments to the list. Contact the moderators at cocoa-dev-admins(at)lists.apple.com Help/Unsubscribe/Update your Subscription: https://lists.apple.com/mailman/options/cocoa-dev/zav%40mac.com This email sent to z...@mac.com ___ Cocoa-dev mailing list (Cocoa-dev@lists.apple.com) Please do not post admin requests or moderator comments to the list. Contact the moderators at cocoa-dev-admins(at)lists.apple.com Help/Unsubscribe/Update your Subscription: https://lists.apple.com/mailman/options/cocoa-dev/archive%40mail-archive.com This email sent to arch...@mail-archive.com
Re: set maximum file size
On 8 Oct 2014, at 21:08, Alex Zavatone z...@mac.com wrote: Why not check the file size before writing? Ah, yes, like a kid who picks up the binoculars fat-end first, I was looking at things the wrong way around (how to limit the size instead of getting the size and doing something with that info). Thanks a lot, that should do it! Best P ___ Cocoa-dev mailing list (Cocoa-dev@lists.apple.com) Please do not post admin requests or moderator comments to the list. Contact the moderators at cocoa-dev-admins(at)lists.apple.com Help/Unsubscribe/Update your Subscription: https://lists.apple.com/mailman/options/cocoa-dev/archive%40mail-archive.com This email sent to arch...@mail-archive.com
Re: Memory Warnings
Once your application is in the background it becomes fair game for being terminated. So even if you're using a very reasonable mount of RAM, if you're in the background you can expect that if some other application needs the memory your application will be terminated. -- David Duncan @ My iPhone On Oct 8, 2014, at 2:06 AM, Gerriet M. Denkmann gerr...@mdenkmann.de wrote: On 8 Oct 2014, at 13:42, Quincey Morris quinceymor...@rivergatesoftware.com wrote: (reposted after editing for length) On Oct 7, 2014, at 22:12 , Gerriet M. Denkmann gerr...@mdenkmann.de wrote: Is there a way for the app to find out how much memory it is currently using? Here’s my take on this, which may be way off base: There’s really no practical way to answer such questions any more. The way memory is managed is so varied, so diffused that even if you could get an answer, it would be as likely to lead you astray as to help you. In practice, I think the question is not how much memory your app is using, but how fast your memory use is increasing over time. If it’s not increasing over time It's not. Checked also with Leaks Instrument. ― then you have a fairly easy evaluation. Your app fits (in the foreground) or it crashes. Your app retreats gracefully (to the background) or it’s terminated. It runs in the foreground. But when in background it sometimes gets terminated. When I have a file of 100 KB and do: myData = [NSData dataWithContentsOfFile:options:error: ]; myBytes = myData.bytes; how much Ram am I now using? 100 KB or 200 KB? I.e. does bytes just return a pointer to some internal structure in NSData? Would options = NSDataReadingMappedIfSafe be of some help? The general answer is that you don’t know how much it’s using. In practice, the answer used to be 100KB (“bytes” returned an interior pointer). But this isn’t always true any more. GCD data blocks, which can be bridged to NSData, can come in pieces, and in that case “bytes” may copy the data into contiguous memory. Mapped data might help in common cases, but how does this save you from “it fits or it crashes” when the file turns out to be on a non-local file system, on which mapping doesn’t work? Memory mapping is more about performance than size, perhaps. I made a small iOS app which has just 3 buttons: Data (loads NSData from file, which resides in the app bundle) Bytes (which uses myData.bytes) Count (which sums all bytes in myData) Result with NSDataReadingMappedIfSafe, Persistent Bytes in Instruments, Allocations: No change with any button (always ca. 3.5 MB). Data button takes 1 msec, Bytes 73 μsec, first Count takes 250 msec longer than subsequent Counts. Without NSDataReadingMappedIfSafe: Data button increases Ram use by 15 MB (about size of file); takes 250 msec. Bytes, Count: no change. Bytes takes 13 μsec. All Counts take same time (ca. 50 msec). So NSDataReadingMappedIfSafe seems to be a good idea (at least in my case of a fixed file in the app bundle). You can easily use a construct like GCD semaphores... Sounds too much work for now. Maybe to be considered later. Thanks for your ideas though. Kind regards, Gerriet. ___ Cocoa-dev mailing list (Cocoa-dev@lists.apple.com) Please do not post admin requests or moderator comments to the list. Contact the moderators at cocoa-dev-admins(at)lists.apple.com Help/Unsubscribe/Update your Subscription: https://lists.apple.com/mailman/options/cocoa-dev/david.duncan%40apple.com This email sent to david.dun...@apple.com ___ Cocoa-dev mailing list (Cocoa-dev@lists.apple.com) Please do not post admin requests or moderator comments to the list. Contact the moderators at cocoa-dev-admins(at)lists.apple.com Help/Unsubscribe/Update your Subscription: https://lists.apple.com/mailman/options/cocoa-dev/archive%40mail-archive.com This email sent to arch...@mail-archive.com
Re: Responsive scrolling: prepareContentInRect: not called
On Oct 8, 2014, at 2:53 AM, Antonio Nunes devli...@sintraworks.com wrote: Turns out the problem is a different, and of course, almost embarrassingly silly one. I misspelled the method name: prepareContentInrect rather than prepareContentInRect. Now that I’ve caught that and corrected it, the method is called. Thanks for you effort anyway. Hmm, this sounds like it would make a great compiler warning. File a Radar, maybe? --Kyle Sluder ___ Cocoa-dev mailing list (Cocoa-dev@lists.apple.com) Please do not post admin requests or moderator comments to the list. Contact the moderators at cocoa-dev-admins(at)lists.apple.com Help/Unsubscribe/Update your Subscription: https://lists.apple.com/mailman/options/cocoa-dev/archive%40mail-archive.com This email sent to arch...@mail-archive.com
Waiting for a future runloop-based callback
I thought I understood runloops pretty well, but I'm running into a situation that's got me stumped. Basically it involves how to wait (block) for a future message-send that will be queued onto the current runloop. So I've started an asynchronous task that runs on a background thread. When the task completes, it's going to notify the originating thread via a call to -performSelector:onThread: In some circumstances the program later decides it need to block the originating thread until the operation completes. So it runs a nested runloop like this: while (!taskIsComplete) [[NSRunLoop currentRunLoop] runMode: @myCustomMode beforeDate: [NSDate distantFuture]]; where taskIsComplete is some flag that will get set by the target method invoked from the task thread. (I use a custom runloop mode because, if I used the default mode, the application would respond to user input and other events while waiting in the loop, resulting in re-entrant calls to the application code, causing all sorts of problems. In fact what I'm actually doing is fixing a bug in the current code where it incorrectly uses the default mode, leading to a crash.) Over on the task thread, when it completes the task it calls: [originator performSelector: @selector(taskDone:) onThread: originatingThread withObject: self waitUntilDone: NO modes: @[@myCustomMode]]; where the -taskDone: method will set that taskIsComplete ivar that the wait code is checking. Unfortunately this doesn't work, because the runloop immediately bails out and returns NO from -runMode:beforeDate:. Apparently this is because there are no input sources registered for myCustomMode. Which is true, but that doesn't mean that no events are going to arrive for that mode! The problem seems to be that this part of the runloop implementation doesn't understand about delayed-perform-in-mode calls. What's the workaround? Should I add some sort of no-op input source to the runloop for my custom mode? That smells very hacky. —Jens ___ Cocoa-dev mailing list (Cocoa-dev@lists.apple.com) Please do not post admin requests or moderator comments to the list. Contact the moderators at cocoa-dev-admins(at)lists.apple.com Help/Unsubscribe/Update your Subscription: https://lists.apple.com/mailman/options/cocoa-dev/archive%40mail-archive.com This email sent to arch...@mail-archive.com
Re: Waiting for a future runloop-based callback
On Oct 8, 2014, at 1:45 PM, Jens Alfke j...@mooseyard.com wrote: In some circumstances the program … runs a nested runloop like this: while (!taskIsComplete) [[NSRunLoop currentRunLoop] runMode: @myCustomMode beforeDate: [NSDate distantFuture]]; where taskIsComplete is some flag that will get set by the target method invoked from the task thread. Over on the task thread, when it completes the task it calls: [originator performSelector: @selector(taskDone:) onThread: originatingThread withObject: self waitUntilDone: NO modes: @[@myCustomMode]]; where the -taskDone: method will set that taskIsComplete ivar that the wait code is checking. Unfortunately this doesn't work, because the runloop immediately bails out and returns NO from -runMode:beforeDate:. Apparently this is because there are no input sources registered for myCustomMode. Which is true, but that doesn't mean that no events are going to arrive for that mode! The problem seems to be that this part of the runloop implementation doesn't understand about delayed-perform-in-mode calls. It's not about understanding perform-selector calls. It's simply that there's no input source at the time you're running the run loop. The fact that there may be input sources in the future doesn't matter. What's the workaround? Should I add some sort of no-op input source to the runloop for my custom mode? Yes. Just add an [NSPort port] to the run loop. That smells very hacky. Yeah, but what are you gonna do? Regards, Ken ___ Cocoa-dev mailing list (Cocoa-dev@lists.apple.com) Please do not post admin requests or moderator comments to the list. Contact the moderators at cocoa-dev-admins(at)lists.apple.com Help/Unsubscribe/Update your Subscription: https://lists.apple.com/mailman/options/cocoa-dev/archive%40mail-archive.com This email sent to arch...@mail-archive.com
Waiting for callback with GCD: impossible?
On a related topic from my last post, looking at the equivalent situation but using GCD instead of runloops, the problem appears to be impossible. That is: * Code running on dispatch queue A starts some async task that will run on another queue B, and which will signal completion by using dispatch_async to call a completion block on queue A, passing it some result. * After the task starts, the logic on queue A decides it needs to block until the task finishes and returns its result. (This might be part of an implementation of the 'future' design pattern, for example.) * Now, how does queue A wait for the callback? It seems to be impossible: the only way for code running on queue A to receive the callback is to return back to the queue dispatcher, so if it tries to wait it simply blocks the queue, causing a deadlock. This is exactly why runloops have the ability to run a nested loop, to block but still allow incoming messages. Of course that comes with its own set of complications, like the necessity for multiple modes, and issues like the one I just brought up. But given that dispatch queues can't do this, what's the solution to this design problem, i.e. turning an async call into a sync one? —Jens ___ Cocoa-dev mailing list (Cocoa-dev@lists.apple.com) Please do not post admin requests or moderator comments to the list. Contact the moderators at cocoa-dev-admins(at)lists.apple.com Help/Unsubscribe/Update your Subscription: https://lists.apple.com/mailman/options/cocoa-dev/archive%40mail-archive.com This email sent to arch...@mail-archive.com
Re: Waiting for callback with GCD: impossible?
* Now, how does queue A wait for the callback? It seems to be impossible: the only way for code running on queue A to receive the callback is to return back to the queue dispatcher, so if it tries to wait it simply blocks the queue, causing a deadlock. You could use some other signaling mechanism -- a dispatch_semaphore_t or NSConditionLock, for example. ___ Cocoa-dev mailing list (Cocoa-dev@lists.apple.com) Please do not post admin requests or moderator comments to the list. Contact the moderators at cocoa-dev-admins(at)lists.apple.com Help/Unsubscribe/Update your Subscription: https://lists.apple.com/mailman/options/cocoa-dev/archive%40mail-archive.com This email sent to arch...@mail-archive.com
Re: Waiting for callback with GCD: impossible?
On Oct 8, 2014, at 11:59 AM, Jens Alfke j...@mooseyard.com wrote: * Now, how does queue A wait for the callback? It seems to be impossible: the only way for code running on queue A to receive the callback is to return back to the queue dispatcher, so if it tries to wait it simply blocks the queue, causing a deadlock. … But given that dispatch queues can't do this, what's the solution to this design problem, i.e. turning an async call into a sync one? Simply performing the rest of the task in the callback is one way. -- Seth Willits ___ Cocoa-dev mailing list (Cocoa-dev@lists.apple.com) Please do not post admin requests or moderator comments to the list. Contact the moderators at cocoa-dev-admins(at)lists.apple.com Help/Unsubscribe/Update your Subscription: https://lists.apple.com/mailman/options/cocoa-dev/archive%40mail-archive.com This email sent to arch...@mail-archive.com
Re: Waiting for callback with GCD: impossible?
On Oct 8, 2014, at 1:59 PM, Jens Alfke j...@mooseyard.com wrote: * Code running on dispatch queue A starts some async task that will run on another queue B, and which will signal completion by using dispatch_async to call a completion block on queue A, passing it some result. * After the task starts, the logic on queue A decides it needs to block until the task finishes and returns its result. (This might be part of an implementation of the 'future' design pattern, for example.) Queues are not like run loops nor threads. Your desire to have ongoing logic occur on a particular queue seems odd to me. The main purpose for custom queues would be to synchronize access to specific resources. Why is there logic on queue A that's doing anything other than briefly accessing a shared resource? For what it's worth, GCD isn't a good fit for the future design pattern. The GCD way of doing this is to have a task that will be executed asynchronously at the time the result becomes available. The A logic shouldn't wait; it should arrange for whatever work it would do when the wait is over to happen at that time. Then it should end. * Now, how does queue A wait for the callback? It seems to be impossible: the only way for code running on queue A to receive the callback is to return back to the queue dispatcher, so if it tries to wait it simply blocks the queue, causing a deadlock. Yup. But given that dispatch queues can't do this, what's the solution to this design problem, i.e. turning an async call into a sync one? Avoid it. The desire to turn async work to synchronous work is the root of the problem. That said, the task on queue B could signal a semaphore or leave a group, rather than attempting to submit a task to queue A. If you must block waiting for the result, the A logic could wait on the semaphore or group. Better would be to use dispatch_group_notify() to just perform the when-the-result-is-ready work when it's ready. Regards, Ken ___ Cocoa-dev mailing list (Cocoa-dev@lists.apple.com) Please do not post admin requests or moderator comments to the list. Contact the moderators at cocoa-dev-admins(at)lists.apple.com Help/Unsubscribe/Update your Subscription: https://lists.apple.com/mailman/options/cocoa-dev/archive%40mail-archive.com This email sent to arch...@mail-archive.com
UIWebView not scaling page (iOS8)
Hi list, I have a UIWebView inside a UITableViewCell that simply won’t scale the page to fit on iOS8. I have set it to scale pages to fit in IB and have also tried setting it in code, but it simply won’t do it. The UIWebView depicts a JPG image located at a web location. I had no trouble with this on iOS7. It may be bad practise to stuff a web view inside a tableViewCell, I am not sure. But it should still be able to scale the page to fit into it, shouldn’t it? Many thanks, Diederik ___ Cocoa-dev mailing list (Cocoa-dev@lists.apple.com) Please do not post admin requests or moderator comments to the list. Contact the moderators at cocoa-dev-admins(at)lists.apple.com Help/Unsubscribe/Update your Subscription: https://lists.apple.com/mailman/options/cocoa-dev/archive%40mail-archive.com This email sent to arch...@mail-archive.com
Re: Waiting for callback with GCD: impossible?
On Oct 8, 2014, at 12:06 PM, davekeck . davek...@gmail.com wrote: You could use some other signaling mechanism -- a dispatch_semaphore_t or NSConditionLock, for example. Yeah, but that doesn't work for the async mode, where I need a dispatch_async call to queue A to notify it. Maybe I'd also need a shared flag that indicates async vs. sync mode, and queue B would look at the flag and either dispatch a block or toggle the condition based on its state. —Jens ___ Cocoa-dev mailing list (Cocoa-dev@lists.apple.com) Please do not post admin requests or moderator comments to the list. Contact the moderators at cocoa-dev-admins(at)lists.apple.com Help/Unsubscribe/Update your Subscription: https://lists.apple.com/mailman/options/cocoa-dev/archive%40mail-archive.com This email sent to arch...@mail-archive.com
Creating a Photo Extension
Nothing leaves you feeling like a stroke victim more than doing something you know you've done before and having it fail all over the place. I'm trying to create a simple application with a dummy photo extension written in swift. I have already written an actual application with a photo extension and that works (as well as it can given the occasional nondeterministic behavior we and others have been seeing). To make this demo app, made a new single-view application in Xcode. Then I add a photo extension target, making sure to select Swift as the language. Then I make sure to go tell the application target that Embedded content contains swift code because otherwise the photo extension will fail at runtime with the *wildly misleading* and unhelpful error message: blah blah blah swiftrelated.dylib blah:image not found (not, in fact, referring to the image the photo extension has been asked to load). Then I delete the Main.storyboard because I don't like storyboards. Then I edit the info.plist created for the extension to remove the NSExtensionMainStoryboard key and replace it with an NSExtensionPrincipalClass key. Then I run the photo extension target in the Photos app. Upon selecting a photo and attempting to edit it with my new extension, I get this error: Terminating app due to uncaught exception 'NSInvalidArgumentException', reason: '*** setObjectForKey: object cannot be nil (key: __NSConcreteUUID 0x7ffdd3f12c60 21074A67-F7EF-423D-A245-8805C0CC134F)' *** First throw call stack: ( 0 CoreFoundation 0x000109d843f5 __exceptionPreprocess + 165 1 libobjc.A.dylib 0x00010b8f1bb7 objc_exception_throw + 45 2 CoreFoundation 0x000109c8af18 -[__NSDictionaryM setObject:forKey:] + 968 3 libextension.dylib 0x00010b8b217a -[_NSExtensionContextVendor _setPrincipalObject:forUUID:] + 114 4 libextension.dylib 0x00010b8b1880 __105-[_NSExtensionContextVendor _beginRequestWithExtensionItems:listenerEndpoint:withContextUUID:completion:]_block_invoke + 822 5 libdispatch.dylib 0x00010cdd0c06 _dispatch_call_block_and_release + 12 6 libdispatch.dylib 0x00010cdeaaf4 _dispatch_client_callout + 8 7 libdispatch.dylib 0x00010cdd62e9 _dispatch_main_queue_callback_4CF + 490 8 CoreFoundation 0x000109cec569 __CFRUNLOOP_IS_SERVICING_THE_MAIN_DISPATCH_QUEUE__ + 9 9 CoreFoundation 0x000109caf46b __CFRunLoopRun + 2043 10 CoreFoundation 0x000109caea06 CFRunLoopRunSpecific + 470 11 GraphicsServices0x00011008a9f0 GSEventRunModal + 161 12 UIKit 0x00010a60e550 UIApplicationMain + 1282 13 libxpc.dylib0x00010d0e3b4b _xpc_objc_main + 416 14 libxpc.dylib0x00010d0e5ee0 xpc_main + 185 15 Foundation 0x00010a3311d1 service_connection_handler + 0 16 PlugInKit 0x00010960fba2 -[PKService run] + 521 17 PlugInKit 0x00010960f867 +[PKService main] + 55 18 PlugInKit 0x00010960fbc6 +[PKService _defaultRun:arguments:] + 17 19 libextension.dylib 0x00010b8bfcdd NSExtensionMain + 51 20 libdyld.dylib 0x00010ce1b145 start + 1 ) libc++abi.dylib: terminating with uncaught exception of type NSException Well, the ONLY reference to this that I can find is that some framework isn't being implicitly loaded and is dying here. See http://www.atomicbird.com/blog/ios-app-extension-tips at the section titled Enabling modules might not work as expected I have literally been pulling my hair out for two hours because I can't get my boilerplate to work and I was actually trying to make a useful demo project, not just bash my head against the same wall (not to mention the hour that I spent trying to get Xcode to NOT build swift successfully when it had copy-pasted Objective-C in it that was clearly not correct [which required me to restart my entire computer]). What am I doing wrong that I need to do right? -- Daniel Blakemore Pixio Software ___ Cocoa-dev mailing list (Cocoa-dev@lists.apple.com) Please do not post admin requests or moderator comments to the list. Contact the moderators at cocoa-dev-admins(at)lists.apple.com Help/Unsubscribe/Update your Subscription: https://lists.apple.com/mailman/options/cocoa-dev/archive%40mail-archive.com This email sent to arch...@mail-archive.com
Re: Waiting for callback with GCD: impossible?
On Oct 8, 2014, at 12:28 PM, Ken Thomases k...@codeweavers.com wrote: Queues are not like run loops nor threads. Your desire to have ongoing logic occur on a particular queue seems odd to me. I don't think it's odd to want an ongoing single flow of control for specific parts of the program. Limiting parallelism makes it a lot easier to reason about the code's behavior. You can create one of those with a background thread, but my understanding is that we're encouraged to use serial dispatch queues instead. I've certainly done this with threads in the past, but there's complexity with setting up the thread's main method that runs a runloop and knows how to exit when it's done; plus, newer system APIs seem to be moving away from runloops (e.g. NSURLSession.) This is part of why I still find the runloop-vs-queue-vs-NSOperation issue such a headache. Apple's given us a lot of concurrency options, and it's really unclear which ones to use when. Especially when a lot of the Cocoa classes work with only some of them but not others. (For example NSStream only supports runloop scheduling while NSURLSession only supports NSOperationQueues.) The main purpose for custom queues would be to synchronize access to specific resources. Why is there logic on queue A that's doing anything other than briefly accessing a shared resource? I disagree about that, actually. The advice to wrap critical sections in dispatch_async calls turns out to produce a _lot_ of overhead. I briefly used it in some performance-sensitive code this spring, but had to rip it out because a large fraction of runtime was being spent in OS code that scheduled dispatch queues, copied/freed blocks, juggled retain counts, etc. It was a lot faster to just use @synchronized, and faster still to make the classes in question single-threaded and push the synchronization tasks up to a higher level of the code. (This isn't just obsessing over clock cycles; I'm literally talking about 2x performance improvements in high-level benchmarks.) —Jens ___ Cocoa-dev mailing list (Cocoa-dev@lists.apple.com) Please do not post admin requests or moderator comments to the list. Contact the moderators at cocoa-dev-admins(at)lists.apple.com Help/Unsubscribe/Update your Subscription: https://lists.apple.com/mailman/options/cocoa-dev/archive%40mail-archive.com This email sent to arch...@mail-archive.com
Get text with hyphenation
Hi, I have to draw several lines of text, one by one, to an HTML page.
I get the text from an NSTextView, iterating throught the text lines:
lineRect = [layoutManager lineFragmentRectForGlyphAtIndex:index
effectiveRange:lineGlyphRange];
index = NSMaxRange(lineGlyphRange);
lineTextRange = [layoutManager
characterRangeForGlyphRange:lineGlyphRange
actualGlyphRange:actualTextRange];
textLine = [self.textStorage attributedSubstringFromRange:lineTextRange]
It works, but textLine doens't contain the hyphenation dash, if visible in
the textView. For example, if the text line ends with last word of the li-
I just get last word of the li.
I have found a dirty workaround
if(lineGlyphRange.length lineTextRange.length)
{
// I manually add a dash - at the end of the textLine extracted.
}
but I guess it will not work in many other cases, e.g. when the glyphs are
more than the chars because of some umlaut...
How to get that dash, if any? Thank you.
Regards
-- Leonardo
___
Cocoa-dev mailing list (Cocoa-dev@lists.apple.com)
Please do not post admin requests or moderator comments to the list.
Contact the moderators at cocoa-dev-admins(at)lists.apple.com
Help/Unsubscribe/Update your Subscription:
https://lists.apple.com/mailman/options/cocoa-dev/archive%40mail-archive.com
This email sent to arch...@mail-archive.com
Re: Waiting for callback with GCD: impossible?
On Oct 8, 2014, at 3:15 PM, Jens Alfke j...@mooseyard.com wrote:
On Oct 8, 2014, at 12:28 PM, Ken Thomases k...@codeweavers.com wrote:
Queues are not like run loops nor threads. Your desire to have ongoing
logic occur on a particular queue seems odd to me.
I don't think it's odd to want an ongoing single flow of control for specific
parts of the program. Limiting parallelism makes it a lot easier to reason
about the code's behavior.
We're not talking about parallelism here, at least not any more so than you
already had by having A and B. We're talking about (a)synchronicity, which is
a separate thing.
For example, this is still serial execution (of this local logic) and a single
flow of control:
mainThreadTask1();
dispatch_async(dispatch_get_global_queue(DISPATCH_QUEUE_PRIORITY_DEFAULT,
0), ^{
backgroundTask1();
dispatch_async(dispatch_get_main_queue(), ^{
mainThreadTask2();
dispatch_async(dispatch_get_global_queue(DISPATCH_QUEUE_PRIORITY_DEFAULT, 0), ^{
backgroundTask2();
dispatch_async(dispatch_get_main_queue(), ^{
mainThreadTask3();
// etc.
});
});
});
});
You can create one of those with a background thread, but my understanding is
that we're encouraged to use serial dispatch queues instead.
I don't know where you got that impression, but that seems like a terrible idea
to me. First of all, a queue is just a data structure. A queue doesn't run
anything, itself. It is still serviced by threads. If you're not using the
data management features of the queue — its width, suspending/resuming, barrier
tasks, etc. — there's little point.
Submitting long-running tasks to a global concurrent queue is a perfectly
acceptable way to get them to run on a background thread. Some people object,
but I have no problem with it. It doesn't have any particular advantage over
spawning a thread, though.
But using a custom queue for that, just to avoid creating a thread explicitly,
seems pointless to me. And I especially don't see the point in using a serial
queue when you don't need to synchronize.
The main purpose for custom queues would be to synchronize access to
specific resources. Why is there logic on queue A that's doing anything
other than briefly accessing a shared resource?
I disagree about that, actually. The advice to wrap critical sections in
dispatch_async calls turns out to produce a _lot_ of overhead. I briefly used
it in some performance-sensitive code this spring, but had to rip it out
because a large fraction of runtime was being spent in OS code that scheduled
dispatch queues, copied/freed blocks, juggled retain counts, etc. It was a
lot faster to just use @synchronized, and faster still to make the classes in
question single-threaded and push the synchronization tasks up to a higher
level of the code. (This isn't just obsessing over clock cycles; I'm
literally talking about 2x performance improvements in high-level benchmarks.)
That doesn't seem like a disagreement to me. I said that's the purpose for
custom queues. You say they aren't well suited for the purpose. That doesn't
change the fact that that's their purpose. If you find them poorly suited,
then that's all the stronger argument not to use them.
Regards,
Ken
___
Cocoa-dev mailing list (Cocoa-dev@lists.apple.com)
Please do not post admin requests or moderator comments to the list.
Contact the moderators at cocoa-dev-admins(at)lists.apple.com
Help/Unsubscribe/Update your Subscription:
https://lists.apple.com/mailman/options/cocoa-dev/archive%40mail-archive.com
This email sent to arch...@mail-archive.com
Table of TextViews
I’m still struggling to create a single scrollview containing a stack of textviews. I created an object called “Subdocument” which has two important readonly properties: NSTextView* myView CGFloat myHeight myHeight returns myView's frame hight, but never less than 10. The Subdocument’s initializers take care of instantiating and connecting up all four text objects (TextView, TextContainer, LayoutManager, and TextStorage). For now, my Document class has one important property which is read/write: NSMutableArray* subdocumentStack; The Document initializer creates and plugs in an array of 20 subdocuments as a test. In my NIB, I add an NSArrayController and set it to manage the Subdocument class and bind it to File’s Owner’s subdocumentStack. Then I add a view-based NSTableView and turn off unnecessary things like column headers, column selection, and column resizing. If I run my application at this point, everything is fine, but the table is empty. NSLog statements confirm the subdocumentStack contains 20 objects with views and heights. Now I try to bind the table. I select the Table View object and bind Table Content to the array controller’s arrangedObjects myView key path. Then I bind its Row Height to the array controller’s arrangedObjects myHeight key path. The next time I run the program, no document ever appears. Initialization happens normally, but before the window opens, an error is logged: [__NSArrayI doubleValue]: unrecognized selector sent to instance address. If I unbind the Row Height, the document window will appear, but the views are all copies of the default Text View Cell. Now, IB liked all the model key paths I chose. subdocumentStack was in the suggestion list for the array controller; and myView and myHeight were in the suggestion lists for table bindings. So here’s my first question: What if anything am I doing wrong when I try to bind the row height? Adding that one thing causes the error and interrupts the NIB from loading. — Charles Jenkins ___ Cocoa-dev mailing list (Cocoa-dev@lists.apple.com) Please do not post admin requests or moderator comments to the list. Contact the moderators at cocoa-dev-admins(at)lists.apple.com Help/Unsubscribe/Update your Subscription: https://lists.apple.com/mailman/options/cocoa-dev/archive%40mail-archive.com This email sent to arch...@mail-archive.com
Re: Table of TextViews
On Oct 8, 2014, at 8:59 PM, Charles Jenkins cejw...@gmail.com wrote: Now I try to bind the table. I select the Table View object and bind Table Content to the array controller’s arrangedObjects myView key path. Then I bind its Row Height to the array controller’s arrangedObjects myHeight key path. The next time I run the program, no document ever appears. Initialization happens normally, but before the window opens, an error is logged: [__NSArrayI doubleValue]: unrecognized selector sent to instance address. What if anything am I doing wrong when I try to bind the row height? Adding that one thing causes the error and interrupts the NIB from loading. The table view's rowHeight binding is not a per-row row height. It's a single number for the height of all rows of the table. Using KVC to get the value from the array controller with the key path arrangedObjects.myHeight returns an array. -valueForKey: applied to an array will always produce an array, and -valueForKeyPath: is just a series of -valueForKey: calls. Then, the table view calls -doubleValue on whatever object KVC obtained. Hence the error that an array class doesn't respond to -doubleValue. If you want separate row heights for each row, you need to implement a table view delegate and the method -tableView:heightOfRow:. You would look up the object for the row by indexing into the array controller's arrangedObjects and obtain the height from your myHeight property. If/when you change the height of a subdocument, you need to inform the table view by calling -noteHeightOfRowsWithIndexesChanged: on it. Regards, Ken ___ Cocoa-dev mailing list (Cocoa-dev@lists.apple.com) Please do not post admin requests or moderator comments to the list. Contact the moderators at cocoa-dev-admins(at)lists.apple.com Help/Unsubscribe/Update your Subscription: https://lists.apple.com/mailman/options/cocoa-dev/archive%40mail-archive.com This email sent to arch...@mail-archive.com
