fix: Potential deadlock when starting the SDK

[philipphofmann]

📜 Description

Fix a deadlock when two threads access SentrySDK.options and SentrySDK.currentHub, which used the same object in synchronized. This problem is fixed now by using two independent locks for SentrySDK.options and SentrySDK.currentHub.

💡 Motivation and Context

Fixes GH-3956, Fixes GH-3899

💚 How did you test it?

Unit tests and as explained in #3956.

📝 Checklist

You have to check all boxes before merging:

• I reviewed the submitted code.
• I added tests to verify the changes.
• No new PII added or SDK only sends newly added PII if sendDefaultPII is enabled.
• I updated the docs if needed.
• Review from the native team if needed.
• No breaking change or entry added to the changelog.
• No breaking change for hybrid SDKs or communicated to hybrid SDKs.

🔮 Next steps

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 90.850%. Comparing base (e072ad1) to head (f3f0c40).

Additional details and impacted files

@@              Coverage Diff              @@
##              main     #3970       +/-   ##
=============================================
+ Coverage   90.844%   90.850%   +0.005%     
=============================================
  Files          593       594        +1     
  Lines        45974     46078      +104     
  Branches     16383     16436       +53     
=============================================
+ Hits         41765     41862       +97     
+ Misses        4139      4036      -103     
- Partials        70       180      +110     

Files	Coverage Δ
Sources/Sentry/SentrySDK.m	86.335% <100.000%> (+0.171%)	⬆️
Tests/SentryTests/SentrySDKTests.swift	95.171% <100.000%> (+0.160%)	⬆️

... and 40 files with indirect coverage changes

---

Continue to review full report in Codecov by Sentry.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update e072ad1...f3f0c40. Read the comment docs.

[github-actions]

Performance metrics 🚀

	Plain	With Sentry	Diff
Startup time	1228.94 ms	1245.88 ms	16.94 ms
Size	21.58 KiB	625.82 KiB	604.24 KiB

Baseline results on branch: main

Startup times

Revision	Plain	With Sentry	Diff
7bb0873	1226.18 ms	1247.30 ms	21.12 ms
dc0db9e	1246.06 ms	1260.46 ms	14.40 ms
26af50b	1223.78 ms	1245.93 ms	22.16 ms
62c15d4	1235.30 ms	1260.82 ms	25.52 ms
b35ccd0	1233.92 ms	1256.69 ms	22.78 ms
3f366ee	1242.28 ms	1260.80 ms	18.52 ms
b2f82fa	1236.94 ms	1262.86 ms	25.92 ms
904d7fa	1225.73 ms	1249.22 ms	23.49 ms
42ef6ba	1201.67 ms	1214.35 ms	12.67 ms
6c31077	1226.24 ms	1237.24 ms	11.00 ms

App size

Revision	Plain	With Sentry	Diff
7bb0873	22.85 KiB	407.09 KiB	384.24 KiB
dc0db9e	20.76 KiB	419.62 KiB	398.86 KiB
26af50b	21.58 KiB	542.38 KiB	520.80 KiB
62c15d4	22.85 KiB	411.14 KiB	388.29 KiB
b35ccd0	21.58 KiB	573.13 KiB	551.55 KiB
3f366ee	20.76 KiB	427.84 KiB	407.08 KiB
b2f82fa	20.76 KiB	419.62 KiB	398.86 KiB
904d7fa	20.76 KiB	432.87 KiB	412.11 KiB
42ef6ba	21.58 KiB	417.86 KiB	396.28 KiB
6c31077	22.84 KiB	401.66 KiB	378.81 KiB

[brustolin]

This probably solve the problem. But Im wondering whether reading the options need to be synchronised. We never change it or one of its properties, and the two functions related to options changed in this PR are internal.

[philipphofmann]

But Im wondering whether reading the options need to be synchronised.

That's an excellent question. We access the options in various places in the SDK, and if users access things like captureMessage, which calls to the currentHub, on a background thread while starting the SDK, we could run into race conditions or other bugs. The overhead should be minimal, so I think it's good to keep the synchronization for now.