fix: Use correct parsing for stackframes

[noahsmartin]

Fixes #4040

The problem was that for some events such as app hangs and CPUDiagnostics the metric kit payload contains an entire flamegraph. This change parses the flame graph into samples, and then picks the longest duration sample as the one we use for a stack trace. It also fixes over-counting the CPUDiagnostics which was emitting more than one event per exception. This is also fixed by identifying the longest duration sample.

Eventually we should support showing the full flame graph on the frontend, but with this PR the bug is fixed and we can consider that a future improvement.

I tested this on test flight by triggering crashes and verifying they are still correct, as well as manually replaying hang reports received in production

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 84.895%. Comparing base (6c20265) to head (ac40ad2).
⚠️ Report is 7 commits behind head on main.
✅ All tests successful. No failed tests found.

Additional details and impacted files

@@              Coverage Diff              @@
##              main     #6908       +/-   ##
=============================================
- Coverage   84.968%   84.895%   -0.073%     
=============================================
  Files          457       459        +2     
  Lines        27608     27575       -33     
  Branches     12142     12145        +3     
=============================================
- Hits         23458     23410       -48     
- Misses        4109      4126       +17     
+ Partials        41        39        -2     

Files with missing lines	Coverage Δ
Sources/Sentry/SentryFormatterSwift.m	100.000% <100.000%> (ø)
Sources/Sentry/SentryMetricKitIntegration.m	100.000% <100.000%> (ø)
...Core/MetricKit/SentryMXCallStackTree+Parsing.swift	100.000% <100.000%> (ø)
...s/Swift/Core/MetricKit/SentryMXCallStackTree.swift	100.000% <ø> (+25.000%)	⬆️
Sources/Swift/Core/MetricKit/SentryMXManager.swift	100.000% <100.000%> (ø)

... and 18 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 6c20265...ac40ad2. Read the comment docs.

[github-actions]

Performance metrics 🚀

	Plain	With Sentry	Diff
Startup time	1192.81 ms	1220.30 ms	27.49 ms
Size	24.14 KiB	1.03 MiB	1.01 MiB

Baseline results on branch: main

Startup times

Revision	Plain	With Sentry	Diff
af848c4	1224.02 ms	1251.00 ms	26.98 ms
2ec2700	1238.28 ms	1258.82 ms	20.53 ms
b358363	1228.33 ms	1262.34 ms	34.01 ms
dcaec4e	1209.06 ms	1246.02 ms	36.96 ms
5ff1b04	1226.00 ms	1259.98 ms	33.98 ms
1b9991e	1233.45 ms	1256.61 ms	23.17 ms
f8cad3c	1217.94 ms	1257.96 ms	40.02 ms
5cfc768	1220.74 ms	1245.06 ms	24.32 ms
650d802	1231.86 ms	1255.64 ms	23.78 ms
5ec90e0	1235.57 ms	1258.45 ms	22.88 ms

App size

Revision	Plain	With Sentry	Diff
af848c4	24.14 KiB	1.01 MiB	1012.94 KiB
2ec2700	23.75 KiB	980.80 KiB	957.05 KiB
b358363	23.75 KiB	987.92 KiB	964.18 KiB
dcaec4e	24.15 KiB	1.01 MiB	1014.80 KiB
5ff1b04	23.75 KiB	1.01 MiB	1016.24 KiB
1b9991e	23.75 KiB	908.01 KiB	884.26 KiB
f8cad3c	23.75 KiB	1.01 MiB	1016.13 KiB
5cfc768	23.75 KiB	850.73 KiB	826.98 KiB
650d802	23.74 KiB	913.13 KiB	889.39 KiB
5ec90e0	23.74 KiB	872.67 KiB	848.92 KiB

Previous results on branch: mxRewriting

Startup times

Revision	Plain	With Sentry	Diff
c20c3c1	1209.08 ms	1237.69 ms	28.61 ms
d0cf095	1225.92 ms	1255.11 ms	29.20 ms
44da6c5	1234.22 ms	1266.02 ms	31.80 ms
bf3e245	1222.17 ms	1257.09 ms	34.92 ms
793ffea	1229.79 ms	1264.65 ms	34.86 ms
c5b3f7a	1220.18 ms	1259.44 ms	39.25 ms
f3b889c	1221.57 ms	1252.77 ms	31.20 ms
8824c94	1226.98 ms	1262.65 ms	35.67 ms

App size

Revision	Plain	With Sentry	Diff
c20c3c1	24.14 KiB	1.03 MiB	1.01 MiB
d0cf095	24.14 KiB	1.03 MiB	1.01 MiB
44da6c5	24.14 KiB	1.03 MiB	1.00 MiB
bf3e245	24.14 KiB	1.03 MiB	1.00 MiB
793ffea	24.14 KiB	1.03 MiB	1.00 MiB
c5b3f7a	24.14 KiB	1.01 MiB	1013.45 KiB
f3b889c	24.14 KiB	1.03 MiB	1.01 MiB
8824c94	24.14 KiB	1.01 MiB	1014.88 KiB

[philipphofmann]

Thank you @noahsmartin, I added a few comments.

[noahsmartin]

Thanks @philipphofmann this is ready for another look!

[philipphofmann]

Thanks, LGTM.

[philipphofmann]

Again, LGTM