dyld crash at launch

Question

dyld crash at launch

alloy opened this issue 8 years ago · comments

A colleague ran into a situation where the app would crash on launch. Important to note is that she was also seeing it with other apps, such as LinkedIn.app.
A restart of her device ‘fixed’ the issue.
These crashes were not being reported to HockeyApp, meaning it was definitely happening very early in the process.
Extracting the crash reports directly from the device revealed a dyld error.
Checking iTunesConnect for crash reports revealed that we had ~80 similar crash reports.

TL;DR

This appears to happen when the device is out of memory, a restart of the device ‘fixes’ it. Presumably this neat trick does so as well.

Info about the errors

Dyld Error Message:
Dyld Message: Library not loaded: @rpath/CocoaLumberjack.framework/CocoaLumberjack
  Referenced from: /var/mobile/Containers/Bundle/Application/D693664B-F6D0-43CB-A839-358E15CAF9E6/Artsy.app/Artsy
  Reason: no suitable image found.  Did find:
    /private/var/mobile/Containers/Bundle/Application/D693664B-F6D0-43CB-A839-358E15CAF9E6/Artsy.app/Frameworks/CocoaLumberjack.framework/CocoaLumberjack:
    mremap_encrypted() => -1, errno=12 for /private/var/mobile/Containers/Bundle/Application/D693664B-F6D0-43CB-A839-358E15CAF9E6/Artsy.app/Frameworks/CocoaLumberjack.framework/CocoaLumberjack

    /private/var/mobile/Containers/Bundle/Application/D693664B-F6D0-43CB-A839-358E15CAF9E6/Artsy.app/Frameworks/CocoaLumberjack.framework/CocoaLumberjack:
    mremap_encrypted() => -1, errno=12 for /private/var/mobile/Containers/Bundle/Application/D693664B-F6D0-43CB-A839-358E15CAF9E6/Artsy.app/Frameworks/CocoaLumberjack.framework/CocoaLumberjack

    /private/var/mobile/Containers/Bundle/Application/D693664B-F6D0-43CB-A839-358E15CAF9E6/Artsy.app/Frameworks/C
  Dyld Version: 370.1

The error message and codes are always the same.
The error is not specific to a framework, it happens with various frameworks.
It appears to have started from iOS 9.1 and up (but that could very well be iOS 9).

Known affected versions

These are the various versions for which we have crash reports.

iOS versions

~/D/dyld crash » ack -h 'OS Version:' **/*.crash | sort | uniq
OS Version:          iOS 9.1 (13B143)
OS Version:          iOS 9.2 (13C75)
OS Version:          iOS 9.2.1 (13D15)
OS Version:          iOS 9.2.1 (13D20)
OS Version:          iOS 9.3 (13E5214d)
OS Version:          iOS 9.3 (13E5225a)

dyld versions

~/D/dyld crash » ack -h 'Dyld Version:' **/*.crash | sort | uniq
  Dyld Version: 370.1
  Dyld Version: 370.6
  Dyld Version: 390.7

Device versions

~/D/dyld crash » ack -h 'Hardware Model:' **/*.crash | sort | uniq
Hardware Model:      iPad3,4
Hardware Model:      iPad4,4
Hardware Model:      iPad4,7
Hardware Model:      iPad5,3
Hardware Model:      iPad5,4
Hardware Model:      iPad6,7
Hardware Model:      iPad6,8
Hardware Model:      iPhone6,1
Hardware Model:      iPhone7,1
Hardware Model:      iPhone7,2
Hardware Model:      iPhone8,1
Hardware Model:      iPhone8,2

Retracing

The error is raised here in dyld.
Only place where mremap_encrypted could return -1 is here.
vn_getpath calls build_path, which in turn has a few possible locations where it could be returning -1:
But trying to hunt that further down would take a lot more time, suffice to say that:
- if mremap_encrypted returns -1, it most likely corresponds to ERESTART: https://github.com/opensource-apple/xnu/blob/2fa84067f6cdeb23267f877ca4fd26201316da1b/bsd/sys/errno.h#L270
- if errno is 12, it most likely corresponds to ENOMEM: https://github.com/opensource-apple/xnu/blob/2fa84067f6cdeb23267f877ca4fd26201316da1b/bsd/sys/errno.h#L102

Caveats

Alas, the latest dyld sources to be released (at the time of writing) appear to be for version 360.18.

Taking a cursory look at the decompiled code for ImageLoaderMachOCompressed::registerEncryption from a firmware image for iOS 9.2 for iPad3,4, does not appear to have any real changes.

~/t/ios-9.2-iPad3,4 » unzip iPad3,4_9.2_13C75_Restore.ipsw 
Archive:  iPad3,4_9.2_13C75_Restore.ipsw
  inflating: 058-25903-078.dmg       
  inflating: 058-26242-078.dmg       
  inflating: 058-26268-078.dmg
[…]

~/t/ios-9.2-iPad3,4 » ~/Code/ObjectiveC/Debugging/xpwn/dmg/dmg extract 058-26268-078.dmg extracted.dmg -k 7ccd940dfda4b304c87050997e9fd70b458f21b5f3845262d985730b23f41e84aa25dae3
Writing out data..
[…]

~/t/ios-9.2-iPad3,4 » open extracted.dmg
~/t/ios-9.2-iPad3,4 » hopper -e /Volumes/Castlerock13C75.P101OS/usr/lib/dyld

int ImageLoaderMachOCompressed::registerEncryption(encryption_info_command const*, ImageLoader::LinkContext const&)(void * arg0, void * arg1) {
    r0 = arg0;
    r7 = (sp - 0x4 - 0x4 - 0x4 - 0x4 - 0x4) + 0xc;
    sp = sp - 0x4 - 0x4 - 0x4 - 0x4 - 0x4 - 0x4 - 0x4 - 0x4 - 0x4;
    r6 = arg1;
    r8 = r2;
    r4 = r0;
    if (r6 != 0x0) {
            r5 = 0x0;
            do {
                    r2 = *(*r4 + 0xe8);
                    if ((r2)(r4, r5, r2) != 0x0) goto (null);
                    r5 = r5 + 0x1;
            } while (true);
            r1 = r5;
            r2 = *(*r4 + 0x104);
            r0 = (r2)(r4, r1, r2);
            r10 = r6 + 0x8;
            asm{ ldm.w      sl, {r1, r5, sl} };
            r2 = *(r8 + 0xb2);
            r8 = *(r0 + 0x4);
            r6 = r0 + r1;
            r11 = *(r0 + 0x8);
            if (r2 != 0x0) {
                    dyld::log("                      0x%08lX->0x%08lX configured for FairPlay decryption\n");
            }
            r0 = _mremap_encrypted();
            r5 = r0;
            if (r5 != 0x0) {
                    r0 = ___error();
                    r3 = *(r4 + 0x4);
                    r2 = *r0;
                    r1 = r5;
                    r0 = dyld::throwf("mremap_encrypted() => %d, errno=%d for %s\n");
                    // …
            }
    }
    return r0;
}

Conclusion

The crash at launch definitely feels like a very bad experience and I feel like iOS should handle this more gracefully. Maybe show an alert? Will have a think about this and possibly file a radar.

I’m not sure what/if we can do about this. We could probably improve it somewhat by trimming some framework fat from the bundle, but I don’t think these are large anyways.

It would be interesting to know if such failures occurred prior to iOS 9 and if they were labelled differently.

Status

A couple of radar tickets have already been filed:
A DTS ticket has been filed http://twitter.com/senior/status/707996353964728321
A guess has been made as to the issue being a jetsam change https://twitter.com/Catfish_Man/status/707955795493687296 (jetsam is the process in which the kernel basically kills apps to free up memory)

Eloy Durán · Answer 1 · Thu Mar 10 2016 22:56:57 GMT+0800 (China Standard Time)

Dynamic library sizes:

$ cd DerivedData/Artsy/Build/Products/Debug-iphoneos/Artsy.app/Frameworks
$ find . -perm +u=x -type f ! -iname '*.strings' ! -iname '*.ttf' ! -iname '*.otf' -print0 | xargs -0 ls -lh | awk '{print $5, $9}' | column -t | gsort -h -r -k1
13M   ./libswiftCore.dylib
1.5M  ./libswiftFoundation.dylib
1.4M  ./HockeySDK_Source.framework/HockeySDK_Source
1.0M  ./ReactiveCocoa.framework/ReactiveCocoa
956K  ./FBSDKCoreKit.framework/FBSDKCoreKit
656K  ./AFNetworking.framework/AFNetworking
395K  ./libswiftCoreGraphics.dylib
392K  ./SDWebImage.framework/SDWebImage
368K  ./CocoaLumberjack.framework/CocoaLumberjack
363K  ./Adjust.framework/Adjust
337K  ./libswiftDarwin.dylib
294K  ./libswiftCoreAudio.dylib
294K  ./Analytics.framework/Analytics
288K  ./FBSDKLoginKit.framework/FBSDKLoginKit
273K  ./Bolts.framework/Bolts
255K  ./libswiftObjectiveC.dylib
248K  ./libswiftUIKit.dylib
248K  ./libswiftCoreMedia.dylib
241K  ./MMMarkdown.framework/MMMarkdown
240K  ./libswiftDispatch.dylib
240K  ./libswiftAVFoundation.dylib
224K  ./Mantle.framework/Mantle
210K  ./ARAnalytics.framework/ARAnalytics
208K  ./libswiftWebKit.dylib
208K  ./libswiftCoreLocation.dylib
208K  ./libswiftCoreData.dylib
208K  ./libswiftContacts.dylib
202K  ./libswiftCoreImage.dylib
163K  ./NAMapKit.framework/NAMapKit
151K  ./AFOAuth1Client.framework/AFOAuth1Client
146K  ./iRate.framework/iRate
145K  ./ALPValidator.framework/ALPValidator
138K  ./ARTiledImageView.framework/ARTiledImageView
136K  ./VCRURLConnection.framework/VCRURLConnection
135K  ./Interstellar.framework/Interstellar
135K  ./ARGenericTableViewController.framework/ARGenericTableViewController
131K  ./ARCollectionViewMasonryLayout.framework/ARCollectionViewMasonryLayout
128K  ./FLKAutoLayout.framework/FLKAutoLayout
118K  ./EDColor.framework/EDColor
111K  ./JSDecoupledAppDelegate.framework/JSDecoupledAppDelegate
109K  ./ObjectiveSugar.framework/ObjectiveSugar
109K  ./Keys.framework/Keys
108K  ./FXBlurView.framework/FXBlurView
106K  ./Artsy_UILabels.framework/Artsy_UILabels
106K  ./Artsy_UIButtons.framework/Artsy_UIButtons
105K  ./KSDeferred.framework/KSDeferred
104K  ./ORStackView.framework/ORStackView
104K  ./Aerodramus.framework/Aerodramus
103K  ./PBDCarouselCollectionViewLayout.framework/PBDCarouselCollectionViewLayout
103K  ./JLRoutes.framework/JLRoutes
101K  ./MARKRangeSlider.framework/MARKRangeSlider
101K  ./JSBadgeView.framework/JSBadgeView
100K  ./ISO8601DateFormatter.framework/ISO8601DateFormatter
99K   ./UICKeyChainStore.framework/UICKeyChainStore
96K   ./RSSwizzle.framework/RSSwizzle
81K   ./ORKeyboardReactingApplication.framework/ORKeyboardReactingApplication
81K   ./ARASCIISwizzle.framework/ARASCIISwizzle
80K   ./UIView_BooleanAnimations.framework/UIView_BooleanAnimations
80K   ./DRKonamiCode.framework/DRKonamiCode
80K   ./AFNetworkActivityLogger.framework/AFNetworkActivityLogger
79K   ./UIAlertView_Blocks.framework/UIAlertView_Blocks
79K   ./Artsy_UIFonts.framework/Artsy_UIFonts
78K   ./NPKeyboardLayoutGuide.framework/NPKeyboardLayoutGuide
76K   ./MultiDelegate.framework/MultiDelegate
76K   ./Artsy_UIColors.framework/Artsy_UIColors
75K   ./Then.framework/Then
74K   ./DHCShakeNotifier.framework/DHCShakeNotifier

Eloy Durán · Answer 2 · Thu Mar 10 2016 23:21:50 GMT+0800 (China Standard Time)

There is very few mentions about this crash that I could find, maybe because people don’t realise it is happening?

Sadly enough, two of those cases appear to get their app rejected, 1 followed up to say they recompiled and resubmitted and it was accepted. I bet that these were just cases of the reviewer’s device running out of memory and that the second case also got accepted after resubmitting.

Cédric Luthi · Answer 3 · Fri Mar 11 2016 05:20:57 GMT+0800 (China Standard Time)

@JaviSoto had a discussion on twitter last week about the same issue but it looks like no solution surfaced. 🙁

Eloy Durán · Answer 4 · Fri Mar 11 2016 05:38:23 GMT+0800 (China Standard Time)

@0xced Aye, he has let me know, it also sparked others chiming in. I have added a ‘Status’ section to the bottom of the OP.

Dan · Answer 5 · Tue Mar 15 2016 15:38:08 GMT+0800 (China Standard Time)

The jetsam/memory pressure process is described here for reference: http://newosxbook.com/articles/MemoryPressure.html

If the issue can be reliably reproduced it may be worth testing explicit weak linking of some frameworks as described in SDK based development

Eloy Durán · Answer 6 · Fri Apr 01 2016 19:04:10 GMT+0800 (China Standard Time)

My radar has been closed as a duplicate of 24278648.

Eloy Durán · Answer 7 · Fri Apr 01 2016 19:05:16 GMT+0800 (China Standard Time)

@robinsenior Did you get any updates on your DTS?

Robin Senior · Answer 8 · Fri Apr 01 2016 20:27:59 GMT+0800 (China Standard Time)

They initially told me it was fixed in 9.3, but an Apple dev on Twitter told me otherwise. I see crash logs for it in beta 7 but not the full release (AFAIK, I haven't checked in a few days).

Eloy Durán · Answer 9 · Fri Apr 01 2016 22:59:52 GMT+0800 (China Standard Time)

@robinsenior I have crash reports for 9.3 builds 13E233 and 13E234, so definitely not fixed.

If you don’t have any atm and want to follow-up with DTS I can send you some.

Robin Senior · Answer 10 · Fri Apr 01 2016 23:17:11 GMT+0800 (China Standard Time)

@alloy thanks; I just checked and we have some with those build numbers so I'll email them to DTS.

As an aside, my phone is much faster under 9.3. The RAM clearing trick used to be essential just to get apps to launch quickly but now is unnecessary. Something related to reclaiming memory must have been fixed.

Eloy Durán · Answer 11 · Fri Apr 01 2016 23:37:20 GMT+0800 (China Standard Time)

Ace 👍

As an aside, my phone is much faster under 9.3. The RAM clearing trick used to be essential just to get apps to launch quickly but now is unnecessary. Something related to reclaiming memory must have been fixed.

That’s great news! Might update my personal device sooner after all 😉

Lukas Spieß · Answer 12 · Fri Apr 01 2016 23:41:25 GMT+0800 (China Standard Time)

Apparently there have been some changes to dyld in 9.3 which resulted in significant performance improvements: https://github.com/stepanhruda/dyld-image-loading-performance
Didn't fix the crash from this issue though 😕

Eloy Durán · Answer 13 · Fri Apr 01 2016 23:43:24 GMT+0800 (China Standard Time)

Yeah, that’s about #586.

Robin Senior · Answer 14 · Tue Apr 05 2016 03:00:02 GMT+0800 (China Standard Time)

Just got this update from DTS:

We’ve made some progress on this one. It seems to be related to the number of embedded frameworks an app has. I noticed that your app has 40 frameworks (12 of which are the Swift dylibs). You could reduce the chance of your app hitting this bug by combining some of those frameworks. Even cutting the number of frameworks in half would make a difference.

/facepalm

Eloy Durán · Answer 15 · Tue Apr 05 2016 17:10:17 GMT+0800 (China Standard Time)

Oh jeez… So their progress is basically all of what this ticket already describes ¯_(ツ)_/¯

I do get that they want to give you something that’s actionable for you right now, but it would have been nice to know a bit more what their status is on the real issue. (rdar://24278648, which is the one mine got closed in favour of.)

Christopher Tava · Answer 16 · Fri Apr 08 2016 00:57:27 GMT+0800 (China Standard Time)

Hi Alloy - Have you heard anything more about iOS dynamic library loading crasher? or are you guys making any changes to combine frameworks etc?
Thanks in advance for any insight.
-chris

Eloy Durán · Answer 17 · Fri Apr 08 2016 01:07:04 GMT+0800 (China Standard Time)

@ctava Nope, if/when we would hear more, or decide a different course of action, it all goes public in this ticket. I don’t think combining frameworks would help, as it would probably still run out of memory and die in the same way. The only possibility to reduce (not fully get rid of) these issues would be to get rid of ‘fat’ (number of larger frameworks), which is not something we will be doing at this time.

To summarise, imo it’s clearly a bug in iOS, and the radar that mine was closed in favour of is still open, so they haven’t ruled it as “works as intended” yet. Until that time, I don’t think the number of crashes vs the number of users warrants a lot of work on our part. YMMV.

Eloy Durán · Answer 18 · Fri Apr 08 2016 22:28:42 GMT+0800 (China Standard Time)

@robinsenior Thinking about that answer a bit more, could you ask them if they do in fact mean the number of libs vs the size of those libs? If they are positive that reducing the number of libs, regardless of size, is a way to workaround the issue, then that would definitely be helpful info.

Robin Senior · Answer 19 · Fri Apr 08 2016 22:32:12 GMT+0800 (China Standard Time)

@alloy done. I'll keep you posted.

Eloy Durán · Answer 20 · Fri Apr 08 2016 22:37:58 GMT+0800 (China Standard Time)

Ace, thanks 👍

Max Meyers · Answer 21 · Sat Apr 09 2016 01:00:46 GMT+0800 (China Standard Time)

For another data point: we're seeing pretty commonly in our app as well since we switched to using dynamic frameworks with CocoaPods. We have 41 frameworks averaging about 1.5MB each. It's a pretty significant issue for us, so we're going back to a static library in our next release.

Eloy Durán · Answer 22 · Sat Apr 09 2016 03:29:49 GMT+0800 (China Standard Time)

@maxmeyers Are you using Swift yet?

Max Meyers · Answer 23 · Sat Apr 09 2016 03:37:09 GMT+0800 (China Standard Time)

@alloy Fortunately not in production yet.

Eloy Durán · Answer 24 · Sat Apr 09 2016 03:38:11 GMT+0800 (China Standard Time)

@maxmeyers Ok cool. Just wanted to check if the Swift dylibs were not causing this issue for you, but that explains it then. Thanks for the feedback 👍

Robin Senior · Answer 25 · Wed Apr 13 2016 23:17:00 GMT+0800 (China Standard Time)

@alloy DTS says it's the number of libs, not the size.

Eloy Durán · Answer 26 · Wed Apr 13 2016 23:30:55 GMT+0800 (China Standard Time)

@robinsenior Well… that’s unexpected. Are you going to spend time on reducing the number of dynamic libs?

Robin Senior · Answer 27 · Wed Apr 13 2016 23:34:19 GMT+0800 (China Standard Time)

@alloy I'd rather not; our build system is already pretty convoluted due to shared frameworks across multiple projects, primarily built using Cocoapods. Do you know of a simple approach for mashing everything together?

Eloy Durán · Answer 28 · Wed Apr 13 2016 23:35:41 GMT+0800 (China Standard Time)

@robinsenior I assume you’re using frameworks because you have some libs that use Swift? Do you have many of those vs objc ones?

Robin Senior · Answer 29 · Fri Apr 15 2016 04:39:19 GMT+0800 (China Standard Time)

@alloy of our 39 frameworks:

12 are libswift
17 are external dependencies
10 are internal frameworks, 7 of which contain Swift

Darryl Bayliss · Answer 30 · Fri Apr 15 2016 06:14:03 GMT+0800 (China Standard Time)

@alloy Do you think it would be possible to trigger this issue even after a device reboot? Had this issue come to light today in our latest app update with a colleague "fixing it" by restarting their device but another still appearing to run into problems on launch.

Eloy Durán · Answer 31 · Fri Apr 15 2016 17:01:11 GMT+0800 (China Standard Time)

@robinsenior Something like https://github.com/johnno1962/Accelerator is a ‘simple’ approach. Alternatively you could maybe have a separate Podfile for your objc pods and statically build those, you might need to do integration into your Xcode project yourself then, though. But make sure that the libs you would be doing this with will be able to lookup their resources, if they have any (wrt to NSBundle lookup).

If/when I decided to do this for our app, I’ll post a follow up with my steps.

Eloy Durán · Answer 32 · Fri Apr 15 2016 17:02:39 GMT+0800 (China Standard Time)

@DarrylBayliss Seeing as DTS is reporting that it has to do with the number of libs, it might well be. Have you retrieved the crash reports from the device after that crash to verify it’s the same?

Orta Therox · Answer 33 · Mon Apr 18 2016 23:01:53 GMT+0800 (China Standard Time)

Want to also note a useful tweet about this issue

@alloy We have less than 20 frameworks including the swift ones and we’re still seeing the issue.

Darryl Bayliss · Answer 34 · Mon Apr 18 2016 23:09:01 GMT+0800 (China Standard Time)

@alloy Managed to grab some logs and we actually had two problems. The dyld issue mentioned here and something unrelated we didn't catch. So nothing new to add here.

Ryan Garchinsky · Answer 35 · Fri Apr 22 2016 21:02:50 GMT+0800 (China Standard Time)

Is this still a thing for 9.3.1? I don't see it listed above but I haven't seen anything to confirm its fixed in 9.3.1. Users are experiencing this in our app as well

Eloy Durán · Answer 36 · Fri Apr 22 2016 21:05:46 GMT+0800 (China Standard Time)

@ryang1428 This happens on 9.3.1 as well.

Ryan Garchinsky · Answer 37 · Fri Apr 22 2016 22:29:48 GMT+0800 (China Standard Time)

thanks @alloy .. did this mean to be 9.3.1 in your original post? OS Version: iOS 9.3 (13E5225a) You have 9.3 listed twice I now noticed

Eloy Durán · Answer 38 · Fri Apr 22 2016 23:10:00 GMT+0800 (China Standard Time)

Nope, 13E5214d & 13E5225a are both 9.3 betas iirc

Victor Ilyukevich · Answer 39 · Wed Apr 27 2016 02:19:54 GMT+0800 (China Standard Time)

@alloy I would like to thank you so much for doing such great writeup!! And also thanks Artsy for doing open source by default. It's great that you can share all details about an investigation of a weird crash! It helps others a lot!

We have noticed same issue with our app. I wonder if anybody is going to bring this topic to the labs at WWDC.

Eloy Durán · Answer 40 · Wed Apr 27 2016 17:50:37 GMT+0800 (China Standard Time)

@yas375 You’re most welcome 🙇

That’s a good question. I’m personally not going, but maybe @orta is?

Robin Senior · Answer 41 · Wed Apr 27 2016 21:38:24 GMT+0800 (China Standard Time)

Not to beat a dead horse, but I've had this crash happen to me in several apps lately (including big ones like LinkedIn). Our app is getting pounded in app store reviews because of it. How is there not a bigger stink about it?

Perry Zheng · Answer 42 · Thu Apr 28 2016 00:43:37 GMT+0800 (China Standard Time)

Just want to chime in on the conversation here. I ran into this issue this morning - the app crashed during startup. I reinstalled the app from app store and still to no avail. And then I found some other apps (not all) are having the same issue. Restarting my phone 'fixed' it.

Keith Smiley · Answer 43 · Thu Apr 28 2016 00:50:34 GMT+0800 (China Standard Time)

Having looked at our crash reports in Xcode we do have a huge number of these crashes.

Joe Fabisevich · Answer 44 · Sat Apr 30 2016 09:20:44 GMT+0800 (China Standard Time)

@robinsenior We're also getting destroyed, hundreds of 1 star reviews. 😢 I've been raising a fuss to everyone I know who could potentially help, and to no avail. It seems like many people don't realize they have these crashes. Since it's a crash on launch, their crash reporting tool (e.g. Crashlytics/Hockey) won't pick it up, and only iTunes Connect can relay it, they're probably living in ignorant bliss.

Brian Michel · Answer 45 · Mon May 02 2016 05:23:14 GMT+0800 (China Standard Time)

Hey folks, talked a bit to a DTS engineer that asked me to keep an eye out for this crash in iOS 9.3.2 beta 3. Might be nothing, but might as well add information here where I can.

Our app was also rejected from app review briefly for this specific problem which involved me coordinating between DTS engineering and the App Review team. Thankfully, we're approved and ready to go now, but things are getting hairy!

P.S. I'll buy whoever is at WWDC with this stack trace a beer as sympathy :)

Eloy Durán · Answer 46 · Mon May 02 2016 06:16:30 GMT+0800 (China Standard Time)

@brianmichel @JaviSoto just mentioned that he’s still seeing it on 9.3.2: http://twitter.com/Javi/status/726882761106038784.

Thanks for putting it on the radar of the App Review team 👍

Eloy Durán · Answer 47 · Tue May 03 2016 03:00:24 GMT+0800 (China Standard Time)

As of now, both @JaviSoto and I have only seen crashes up to 9.3.2 beta 1 (13F51a). This combined with @brianmichel’s DTS feedback to lookout for this issue in beta 3 seems very hopeful.

Christopher Tava · Answer 48 · Tue May 10 2016 00:14:21 GMT+0800 (China Standard Time)

any update on this issue. for example, how hopeful is 9.3.2 beta in terms of fixing the iOS crasher?

Mitchell Porter · Answer 49 · Thu May 12 2016 01:58:46 GMT+0800 (China Standard Time)

We were just rejected for this issue this morning. App crashed on launch for reviewer, looked at the attached crash file, sure enough it was this issue. I didn't even bother with the appeal process, just submitted another build. Hopefully it gets approved.

Ironically, this is our first denial because of this issue, and I had just gotten done uninstalling 4 cocoapods we didn't absolutely need anymore.

This is really starting to worry me, and i'm wondering what percentage of our users are experiencing this and simply uninstalling the app and/or leaving bad reviews. We've already received a few 1 star reviews because of this.

Josh Garnham · Answer 50 · Thu May 12 2016 01:59:53 GMT+0800 (China Standard Time)

We're seeing a number of 1 star reviews because of this too, so much so that in the next release notes we're going to have to specifically attribute this to being Apple/iOS's issue and not ours.

Really do hope that this is fixed in 9.3.2. If not it's definitely going to be something to mention at WWDC labs.

Max Meyers · Answer 51 · Thu May 12 2016 02:04:15 GMT+0800 (China Standard Time)

Here's what happened when we switched back from frameworks to static libraries (this is crashes as reported by iTunes Connect App Analytics):

Josh Garnham · Answer 52 · Thu May 12 2016 02:10:34 GMT+0800 (China Standard Time)

@maxmeyers Did you not require frameworks for use with Swift or did you make use of https://github.com/johnno1962/Accelerator?

Max Meyers · Answer 53 · Thu May 12 2016 02:13:11 GMT+0800 (China Standard Time)

We didn't have any Swift dependencies. One option might be to install your Swift frameworks with Carthage and use CocoaPods for Objective-C dependencies (that's assuming you're currently using CocoaPods to install your frameworks).

Eloy Durán · Answer 54 · Thu May 12 2016 02:15:12 GMT+0800 (China Standard Time)

@maxmeyers Do you use Swift in your app? If so, are you still getting this crash report (albeit in low numbers) due to libSwiftCore.dylib etc?

Max Meyers · Answer 55 · Thu May 12 2016 02:17:46 GMT+0800 (China Standard Time)

No Swift at all. This crash has completely disappeared.

Eloy Durán · Answer 56 · Thu May 12 2016 02:18:13 GMT+0800 (China Standard Time)

One option might be to install your Swift frameworks with Carthage and use CocoaPods for Objective-C dependencies.

Or you could use a second Podfile for your Swift dependencies and use something like Rome, if you prefer CocoaPods ;)

Eloy Durán · Answer 57 · Thu May 12 2016 02:19:08 GMT+0800 (China Standard Time)

@maxmeyers Ah, bugger, had hoped for an extra convincing case to Apple that shows that just using Swift leads to this issue. But great for you of course!

Josh Garnham · Answer 58 · Thu May 12 2016 02:29:35 GMT+0800 (China Standard Time)

@alloy: That wouldn't solve the issue though would it? Only reduce the likelihood of the crash because of less frameworks. Either way, unfortunately our app is 50% Swift so most of our pods are used in both 😞

Eloy Durán · Answer 59 · Thu May 12 2016 02:35:56 GMT+0800 (China Standard Time)

@SquaredTiki Correct.

Dan · Answer 60 · Thu May 12 2016 03:18:45 GMT+0800 (China Standard Time)

Have any of you attempted to weak link frameworks and load them after launch using the NSBundle API?

Orta Therox · Answer 61 · Thu May 12 2016 03:20:31 GMT+0800 (China Standard Time)

With that technique, it sounds like you're starting to move into some of the territory covered in this issue: #586

Dan · Answer 62 · Thu May 12 2016 03:28:53 GMT+0800 (China Standard Time)

There are two, related issues at work here, both centering on the dynamic loader. Memory pressure and CPU. In the crash logs that I have seen memory pressure is the overriding concern. As the loader loads strongly linked frameworks at startup it has to verify their signatures. System processes involved in that increase memory pressure when it may have already been high on the device. The loader can't get enough memory to continue loading frameworks and their dependencies, the application can't be launched or is immediately terminated.

Weak-linking the frameworks and loading after launch using the NSBundle API gives the developer more control over this process.

Monitoring the performance of signature verification may also be helpful.

Jason Patterson · Answer 63 · Fri May 13 2016 20:19:54 GMT+0800 (China Standard Time)

Apple has told us that weak linking does not help the problem.

Thomas Abend · Answer 64 · Sat May 14 2016 02:49:58 GMT+0800 (China Standard Time)

We're having a similar problem and just wanted to reiterate the thanks for Artsy's open source ethos. This thread is very helpful.

freshnitesh · Answer 65 · Sat May 14 2016 03:39:59 GMT+0800 (China Standard Time)

In order to reproduce this crash consistently, we downloaded a number of apps from the App Store on an iOS 9.3.1 device (iPhone 6+). I opened 14 apps, and when trying to open the 15th app, I experienced a crash on launch. From there, any app I tried to open for the first time would crash on launch, while previously opened apps would continue to launch successfully.

I was able to reproduce this same sequence both after a reboot, and after manually killing all open apps via the app switcher.

After upgrading the phone iOS version to 9.3.2 Beta 4, I was unable to reproduce any crash. At the maximum, I opened 50+ apps successfully. Hopefully this means the issue is fixed or at least ameliorated with iOS 9.3.2 Beta 4.

Eloy Durán · Answer 66 · Sat May 14 2016 04:10:25 GMT+0800 (China Standard Time)

@freshnitesh That’s great to hear, tanks for the update!

freshnitesh · Answer 67 · Sat May 14 2016 04:26:14 GMT+0800 (China Standard Time)

@alloy Happy to help :)

UnsafePointer<T> · Answer 68 · Mon May 16 2016 00:42:20 GMT+0800 (China Standard Time)

@freshnitesh thank you so much for the detailed instructions on how to reproduce the crash consistently. I'm also able to confirm that iOS 9.3.2 beta 4 fixes this issue. At some point, apps take longer to launch but no crash so far, which is way better.

Joe Fabisevich · Answer 69 · Tue May 17 2016 04:33:40 GMT+0800 (China Standard Time)

9.3.2 is out. So far nothing is crashing, but will keep an eye on this in the coming days/weeks. 🙏 for success.

Greg Parker · Answer 70 · Thu May 19 2016 10:11:42 GMT+0800 (China Standard Time)

Yes, we believe the crashes on launch with mremap_encrypted -1 errno=12 are fixed in iOS 9.3.2 (as of beta 3, I think). Please file new bug reports if you see it again.

Orta Therox · Answer 71 · Thu May 19 2016 11:49:07 GMT+0800 (China Standard Time)

Alright, I'm calling this 👏

👍

Eloy Durán · Answer 72 · Thu May 19 2016 16:53:43 GMT+0800 (China Standard Time)

Awesome! 👏

Christoffer Winterkvist · Answer 73 · Sat May 21 2016 16:36:26 GMT+0800 (China Standard Time)

This is great! 🚀

Raimon Lapuente · Answer 74 · Fri Sep 23 2016 16:49:56 GMT+0800 (China Standard Time)

That's an amazing thread.
Despite the fact that is fixed in 9.3.2 (and I hope iOS 10), there's still the issue we're supporting quite a lot of old version devices. What you do on those cases?

Orta Therox · Answer 75 · Fri Sep 23 2016 17:33:45 GMT+0800 (China Standard Time)

Aside from the manual work shown in:

A new CocoaPods plugin has been created by @Ruenzuo since then that will automate most of that work:

https://github.com/Ruenzuo/cocoapods-amimono

Other than that, there's no real choice to to write it off. At least 9.3 came with awesome emoji, and that got most of our users to update.

Aagha Abdul Waheed · Answer 76 · Mon Sep 24 2018 06:48:34 GMT+0800 (China Standard Time)

I am getting this error. App is working fine on simulator, but as I run it on device it crashes .I don't know why it's happening.