Wrong value type of language parameter of SDLRegisterAppInterfaceResponse

[mvyrko]

Bug Report

In function - (void)didEnterStateRegistered when we try to check condition [actualLanguage isEqualToEnum:desiredLanguage] we crash with [__NSCFNumber isEqualToEnum:]: unrecognized selector sent to instance

Expected Behavior

No crash

Observed Behavior

Crash

OS & Version Information

• iOS Version: Any
• SDL iOS Version: 6.1.0

Test Case, Sample Code, and / or Example App

[NicoleYarroch]

@mvyrko It would be really useful to see the RegisterAppInterface response being returned by the head unit.
From the bugs you have reported, it looks like the head unit manufacturer is returning unexpected values:

1. Returning the value of null when the parameter should be omitted.
2. Returning a number instead of a string (I assume they are returning 0 or -1 instead of just leaving it blank?).
3. Returning an array when it should be a dictionary. I'm not sure if the array or dictionary has invalid values or isn't formatted correctly.

This is what I am assuming from the bugs you are reporting. Like I said, if you could set a breakpoint and get the actual RegisterAppInterface response returned, it would greatly help with the debugging. Also, in other areas where it is crashing, it would be useful to see the actual response returned from SDL Core.

[mvyrko]

Hello @NicoleYarroch,
Unfortunately, I can't provide these information. I've got crashlogs from fabric, and we have no additional analytics information about head unit models.

[NicoleYarroch]

@mvyrko Is this true of all the bugs you are reporting or just this particular bug (#1159)

[mvyrko]

@NicoleYarroch Yes, It's true of all the bugs

[joeygrover]

I believe the issue #1127 is releated with the same type of issue.

[ghost]

@joeygrover kind of. This issue is about the root cause of #1127. The root cause wasn't fixed but the library was changed in the way to not perform a manager configuration update.

[joeygrover]

@kshala-ford right, I'm adding for visibility that the issue is related to unexpected data being sent from core and causing a crash in the library. The solution was a band aid fix on the single parameter rather than a library wide solution as seen in the check that was added [actualLanguage isKindOfClass:[NSString class]]. So it is the same root cause issue that should be addressed library wide rather than adhoc as the PR #1128 was to fix #1127