`EXC_BAD_ACCESS` Crash in `CoreLogger.debug()`

[msrutek-paylocity]

Stack trace

Thread 1 Queue : com.apple.main-thread (serial)
#0	0x0000000000000000 in 0x00000000 ()
#1	0x000000010cf02210 in InternalLogger.log(_:message:error:) at dd-sdk-ios/DatadogInternal/Sources/Telemetry/InternalLogger.swift:38
#2	0x000000010cf02ee8 in protocol witness for CoreLogger.log(_:message:error:) in conformance InternalLogger ()
#3	0x000000010cf014dc in CoreLogger.debug(_:error:) at dd-sdk-ios/DatadogInternal/Sources/Telemetry/CoreLogger.swift:75
#4	0x000000010cc01b34 in closure #2 in DatadogNTPDateProvider.synchronize(update:) at dd-sdk-ios/DatadogCore/Sources/Core/Context/ServerOffsetPublisher.swift:49
#5	0x000000010cc58f2c in closure #1 in KronosClock.sync(from:samples:first:completion:) at dd-sdk-ios/DatadogCore/Sources/Kronos/KronosClock.swift:142
#6	0x000000010cc59078 in partial apply for closure #1 in KronosClock.sync(from:samples:first:completion:) ()
#7	0x000000010cc61e30 in $defer #1 () in closure #1 in closure #1 in KronosNTPClient.query(pool:version:port:numberOfSamples:maximumServers:timeout:progress:) at dd-sdk-ios/DatadogCore/Sources/Kronos/KronosNTPClient.swift:53
#8	0x000000010cc61930 in closure #1 in closure #1 in KronosNTPClient.query(pool:version:port:numberOfSamples:maximumServers:timeout:progress:) ()
#9	0x000000010cc650b4 in partial apply for closure #1 in closure #1 in KronosNTPClient.query(pool:version:port:numberOfSamples:maximumServers:timeout:progress:) ()
#10	0x000000010cc62b8c in closure #1 in KronosNTPClient.query(ip:port:version:timeout:numberOfSamples:completion:) at dd-sdk-ios/DatadogCore/Sources/Kronos/KronosNTPClient.swift:111
#11	0x000000010cc62d50 in partial apply for closure #1 in KronosNTPClient.query(ip:port:version:timeout:numberOfSamples:completion:) ()
#12	0x000000010cc62fa0 in thunk for @escaping @callee_guaranteed (@guaranteed Data?, @unowned Double) -> () ()
#13	0x000000010cc63554 in closure #2 in KronosNTPClient.query(ip:port:version:timeout:numberOfSamples:completion:) at dd-sdk-ios/DatadogCore/Sources/Kronos/KronosNTPClient.swift:125
#14	0x000000010cc60b20 in static KronosBlockTimer.invokeFrom(timer:) at dd-sdk-ios/DatadogCore/Sources/Kronos/KronosNSTimer+ClosureKit.swift:51
#15	0x000000010cc60b98 in @objc static KronosBlockTimer.invokeFrom(timer:) ()
#16	0x000000018111a8c4 in __NSFireTimer ()

Reproduction steps

We're able to reproduce this in unit tests, when running ~1000 iterations

🖼️

Volume

0.001

Affected SDK versions

2.30.0

Latest working SDK version

Does the crash manifest in the latest SDK version?

Yes

Deployment Target

iOS 15

Device Information

No response

Other relevant information

No response

[ncreated]

Thanks for the report, @msrutek-paylocity, and for including the stack trace! We've been able to reproduce the crash and will work on a fix.

The root cause is in CoreLogger.debug(), and while KronosBlockTimer appears in the stack trace, it's only the entry point that triggers downstream logging. I’ve updated the issue title to reflect the actual failing call.

In the meantime, a valid mitigation is to disable SDK verbosity by setting:

Datadog.verbosityLevel = nil

This disconnects CoreLogger and should prevent the crash from occurring.

[msrutek-paylocity]

Hello @ncreated , thanks for the quick response and taking the time to go through this.

Inspecting our setup, I can't see

• any usage or setting of Datadog.verbosityLevel,
• any usage of DD_DEBUG flag that might set it,

and it also seems that the default for the property is nil. Furthermore, looking at the value throughout the unit test lifetime, it seems like it is still nil .

Due to these reasons, I don't understand how setting

Datadog.verbosityLevel = nil

could help us here.

Is there something we're missing? Thanks!

[mariedm]

Hi @msrutek-paylocity 👋

I've taken another look at this issue and identified the root cause: a race condition when DD.logger is replaced during SDK initialization. While we had already identified a separate issue with the date formatter used in the Logger (mentioned in @ncreated's previous response), the crash you're experiencing is specifically caused by this race condition.

We have a fix coming up, which will be available in the next release.

Thanks again for reporting this issue!

[msrutek-paylocity]

Hello @mariedm , thanks for looking into this!

[ncreated]

@msrutek-paylocity this issue is fixed in 3.2.0, thanks for reporting 🚀!