When I build my app to my actual iPhone the debug area shows this:
[Allocator] Middle guard protection failed %d
[Allocator] Allocator invalid, falling back to malloc
It shows the 2nd line a total of 30 times. I have no idea what it means or how to fix it. It does not show this when I build to the simulator.
I am having issues with getting state preservation to work using storyboards and restoration ID's and I have a feeling this has something to do with a memory issue so it's dumping my memory and therefore I get no app restoration. Basically, when I go back to my app it shows me the last screen I was on for a second and then goes back to the root page.
Anyway, I'd like to fix this malloc stuff so I can at least rule it out the culprit, plus I don't want to have an issue with memory in general...
I've been googling this for a couple of weeks now too and can't find anything!
Looks like it's an issue with the Crashlytics framework. I have the same issue, and commenting this API call:
[Crashlytics startWithAPIKey:API_KEY];
removes that warning.
It does indeed seem to be an issue at Crashlytics. I know from other threads that they raised the "Allocator invalid..." issue in relation to another Middle error (not Middle guard protection failed %d), which later got marked as fixed.
I fixed this by deleting all the crashlytics stuff and using the new fabric/crashlytics framework. Problem solved.
(Interestingly, I had it only on an iPad, my iPhone 5C made no complaints at all.)
Related
I created a completely new and empty FMX app, tried it using Delphi 10.3.3 and 10.4.0 (patch 3), and when installed on Android 7.1.2 devices (tried two different ones and the official simulator VM), after the splash screen this error message appears:
'java.lang.NullPointerException: Attempt to invoke virtual method 'boolean java.lang.String.contains(java.lang.CharSequence)' on a null object reference'
The error message box flickers quickly (endless loop) and gets slower and slower and after a minute the complete device doesn't react anymore and I can only hard reset it.
I found this thread, so I'm not alone. They say it worked until Delphi 10.1.2. There are also many other similar search results.
Is there any solution? On other Android versions it works well.
Or what can I try to get more information about what causes it?
Do you know a good candidate for an app made with Firemonkey that I could download directly and try out on the device?
Now there is a fix available at quality.embarcadero.com!
Just one line has to be added to 'Androidapi.JNIBridge.pas'. I'm not sure if I should copy it here... Will update this answer when it is released.
I'm having a really weird behaviour with my TWA. When I am launching the app, the address bar is not shown and the app is running in standalone mode as expected.
But when I am switching between Apps (putting the app in the background), and coming back to it, something weird happens: the page reloads, and the address bar is shown at the top of my app. I am not really sure to understand what's happening here. Even more strange, it looks like this weird behaviour is not happening all the time.
Does anybody here encountered similar issues ?
I checked my assetlinks file, it's accessible and valid, app bundle and so on, everything looks fine. The fact that the app is not showing the bar at launch also seems to confirm that the problem is not coming from the certificate or a configuration issue. What else could cause the problem ?
First launch (looking perfect):
Back from background:
We have seen this occasionally as well.
Which version of Chrome Custom Tabs are you using? Seems like most articles/tutorials use e446d08 which is quite outdated. Suggest trying a newer/latest version of it:
https://github.com/GoogleChrome/custom-tabs-client/commits/master
Their last commit comment suggests that they are still testing things:
There's definitely more we can test, but this is a start.
https://chromium.googlesource.com/custom-tabs-client/+/da65829d7d4b80c00809c6c4aa7f61f88fc7ca26
My app (OS X 10.10, Xcode 6.1.1) uses an AVPlayerView to show a video resource bundled inside the app. Normally the video runs as expected. Sometimes--and I haven't figured out a way to reliably reproduce this--the video does not start to play, and the screen is all black. At the same time, a warning pops up in the console:
<<<< Boss >>>> figPlaybackBossPrerollCompleted: unexpected preroll-complete notification
How can I programmatically recover from this? I tried Googling parts of the error description but there was only one result (an email to the Apple Mailing List).
Perhaps a notification my view controller can listen for?
Alternatively, if you know a way I can force this state to reproduce, that would help too, as I could then put a breakpoint in.
UPDATE: I still haven't been able to force this problem to reproduce, but I have seen it a couple of times since the first post, and I can say for sure that it's not related to the NSNotifications AVPlayerItemNewErrorLogEntryNotification, AVPlayerItemPlaybackStalledNotification or AVPlayerItemFailedToPlayToEndTimeNotification. My View Controller is listening for those, and didn't hear them when the bug presented.
Thanks!
I have an (old) audio app that is misbehaving on iOS 5.1.1. It records audio and on older iOS versions (don't know precisely where the "break" is) it would stay "in foreground" while recording, without any nudging.
But on 5.1.1 the app is put into background after two minutes, and then things go sour. Currently (will have to change this, I suppose) the app kills recording when it's backgrounded (and it appears to do this successfully), but it still dies with a trap in the above routine.
Unfortunately, the call stack is empty when this occurs, so there's little clue as to why the app's getting killed, but I gather (just from hints here and there on the web) that the trap occurs because a background app cannot use any UI facilities, and the app must somehow be calling something UI-ish. But I haven't a clue what it might be.
I've worked through most of the notifications, to see if a notification might be lurking in a queue somewhere and doing something, but I've not found anything so far that might be triggering a UI opp.
Any ideas on how to track this down?
Aha!! The app uses an Apple freebee widget known as AQLevelMeter. When recording is stopped, the level meter is also stopped, but the stop code inside AQLevelMeter.mm does not invalidate the timer that's driving the UI updates.
Ive set up in app purchases. Everything works correctly except randomly after a couple purchases i get EXC_BAD_ACCESS on the addPayment: line. This i believe is some sort of memory issues that im having a heck of a time trying to figure out. On inspecting things i noticed another peculiar behavior. In SKPaymentTransactionStateFailed i put an NSLog, This log triggers ~50 times before the alert view for the in app purchase pops up, then it stops and i can continue normally with my purchase. Is this normal? Could this be the reason im getting EXC_BAD_ACCESS? Im using xcode 4 preview 6 and programming for iOS5.