feat: Deserializing Events

[philipphofmann]

📜 Description

Add Decodable implementations via Swift extensions to deserialize SentryEvents.

This is the feature branch PR for deserializing SentryEvents. We're going to open multiple PRs to this one, so we have smaller PRs for quicker reviews.

To minimize the scope the goal is not to use Encodable and keep the existing serialization logic.

Tasks

Preview Give feedback

1. feat: Deserializing SentryFrame #4739
   +0
2. SentryGeo
3. feat: Deserialize User #4732
   +0
4. feat: Deserializing SentryNSError #4743
   +0
5. feat: Deserializing SentryStacktrace #4744
   +0
6. feat: Deserializing SentryMechanismMeta #4745
   +0
7. feat: Deserializing SentryMechanism #4746
   +0
8. feat: Deserialize SentryException #4791
   +0
9. feat: Deserialize SentryThread #4790
   +0
10. feat: Deserialize SentryDebugMeta #4789
    +0
11. feat: Deserialize SentryBreadcrumb #4785
    +0
12. feat: Deserialize SentryMessage #4783
    +0
13. feat: Deserialize SentryRequest #4784
    +0
14. SentryEvent
15. How to handle SentryBreadcrumb.unknown?

💡 Motivation and Context

We need this for app hangs #4261.

💚 How did you test it?

📝 Checklist

You have to check all boxes before merging:

• 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.
• I updated the wizard 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

[github-actions]

	Messages
📖	Do not forget to update Sentry-docs with your feature once the pull request gets approved.

Generated by 🚫 dangerJS against 8067173

[philprime]

In case you want to implement a custom encoder for converting to the dictionary, I did implement a custom encoder before in Postie, to parse custom property wrappers/annotations into a structure to send URL requests.

Even though that use case is quite more complicated than just encoding to a dictionary, it might be helpful for reference:

https://github.com/kula-app/Postie/blob/main/Sources/Postie/Encoder/RequestEncoder.swift

[github-actions]

Performance metrics 🚀

	Plain	With Sentry	Diff
Startup time	1223.78 ms	1235.52 ms	11.75 ms
Size	22.31 KiB	805.86 KiB	783.54 KiB

Baseline results on branch: main

Startup times

Revision	Plain	With Sentry	Diff
1a86fb2	1224.69 ms	1236.47 ms	11.77 ms
72c8d84	1218.69 ms	1241.08 ms	22.39 ms
c76f9b0	1249.78 ms	1265.14 ms	15.36 ms
7cd187e	1243.56 ms	1250.20 ms	6.64 ms
ed1c644	1225.92 ms	1241.24 ms	15.33 ms
881a955	1214.27 ms	1235.88 ms	21.61 ms
fb53d97	1235.00 ms	1241.88 ms	6.88 ms
3dcba20	1237.14 ms	1249.69 ms	12.54 ms
b9fc537	1211.00 ms	1220.40 ms	9.40 ms
313b1d9	1240.18 ms	1258.44 ms	18.26 ms

App size

Revision	Plain	With Sentry	Diff
1a86fb2	21.58 KiB	612.11 KiB	590.53 KiB
72c8d84	22.85 KiB	408.88 KiB	386.03 KiB
c76f9b0	22.85 KiB	406.69 KiB	383.84 KiB
7cd187e	20.76 KiB	401.65 KiB	380.89 KiB
ed1c644	21.58 KiB	670.39 KiB	648.81 KiB
881a955	22.85 KiB	407.63 KiB	384.78 KiB
fb53d97	20.76 KiB	425.80 KiB	405.04 KiB
3dcba20	22.31 KiB	780.90 KiB	758.58 KiB
b9fc537	21.58 KiB	676.19 KiB	654.61 KiB
313b1d9	22.85 KiB	414.11 KiB	391.26 KiB

Previous results on branch: feat/deserializing-events

Startup times

Revision	Plain	With Sentry	Diff
242b57f	1236.13 ms	1252.17 ms	16.04 ms
74e111a	1229.31 ms	1244.63 ms	15.33 ms
378d4bd	1216.53 ms	1241.19 ms	24.66 ms
04e8394	1223.24 ms	1244.33 ms	21.08 ms
4dc265c	1221.31 ms	1238.55 ms	17.24 ms
8a562cc	1228.57 ms	1245.18 ms	16.60 ms
8dae1f6	1212.60 ms	1226.52 ms	13.93 ms

App size

Revision	Plain	With Sentry	Diff
242b57f	22.31 KiB	801.48 KiB	779.17 KiB
74e111a	22.31 KiB	800.12 KiB	777.80 KiB
378d4bd	22.31 KiB	784.05 KiB	761.73 KiB
04e8394	22.31 KiB	771.93 KiB	749.62 KiB
4dc265c	22.31 KiB	798.21 KiB	775.90 KiB
8a562cc	22.31 KiB	773.33 KiB	751.02 KiB
8dae1f6	22.31 KiB	772.51 KiB	750.20 KiB

[philipphofmann]

In case you want to implement a custom encoder for converting to the dictionary

Thanks for pointing that out. I would rather spend the time converting all the serialization to Decoable, because that's what we should do on the long run, IMO.

@philprime, does the overall approach look good to you?

[philprime]

Looks good to me.

[philipphofmann]

FYI, @philprime, I ditched the Encodable because we have some tests validating the data for envelopes that break when using Encodable. Switching the whole serialization to Encodable will make fixing the tests easier. I want to keep the scope small for now.

[codecov]

Codecov Report

Attention: Patch coverage is 99.61832% with 5 lines in your changes missing coverage. Please review.

Project coverage is 91.436%. Comparing base (c6a8035) to head (8067173).
Report is 1 commits behind head on main.

Files with missing lines	Patch %	Lines
...s/Swift/Protocol/Codable/DecodeArbitraryData.swift	92.957%	5 Missing ⚠️

Additional details and impacted files

@@              Coverage Diff              @@
##              main     #4724       +/-   ##
=============================================
+ Coverage   91.398%   91.436%   +0.038%     
=============================================
  Files          627       645       +18     
  Lines        74571     75709     +1138     
  Branches     26810     26517      -293     
=============================================
+ Hits         68157     69226     +1069     
- Misses        6315      6390       +75     
+ Partials        99        93        -6     

Files with missing lines	Coverage Δ
Sources/Sentry/SentryGeo.m	95.000% <100.000%> (+1.000%)	⬆️
Sources/Sentry/SentryLevelHelper.m	100.000% <100.000%> (ø)
...ft/Protocol/Codable/NSNumberDecodableWrapper.swift	100.000% <100.000%> (ø)
...ift/Protocol/Codable/SentryBreadcrumbCodable.swift	100.000% <100.000%> (ø)
Sources/Swift/Protocol/Codable/SentryCodable.swift	100.000% <100.000%> (ø)
...wift/Protocol/Codable/SentryDebugMetaCodable.swift	100.000% <100.000%> (ø)
...wift/Protocol/Codable/SentryExceptionCodable.swift	100.000% <100.000%> (ø)
...es/Swift/Protocol/Codable/SentryFrameCodable.swift	100.000% <100.000%> (ø)
...rces/Swift/Protocol/Codable/SentryGeoCodable.swift	100.000% <100.000%> (ø)
...wift/Protocol/Codable/SentryMechanismCodable.swift	100.000% <100.000%> (ø)
... and 24 more

... and 42 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 c6a8035...8067173. Read the comment docs.

[brustolin]

l:

Suggested change

	func decodeFromJSONData<T: Decodable>(jsonData: Data) -> T? {
	if jsonData.isEmpty {
	return nil
	}
	do {
	let decoder = JSONDecoder()
	return try decoder.decode(T.self, from: jsonData)
	} catch {
	SentryLog.error("Could not decode object of type \(T.self) from JSON data due to error: \(error)")
	return nil
	}
	}
	extension Decodable {
	init?(jsonData: Data) {
	guard !jsonData.isEmpty else {
	return nil
	}
	do {
	let decoder = JSONDecoder()
	self = try decoder.decode(Self.self, from: jsonData)
	} catch {
	print("Could not decode object of type \(Self.self) from JSON data due to error: \(error)")
	return nil
	}
	}
	}

Then, instead of

let decoded = decodeFromJSONData(jsonData: data) as Geo?

We have

let decoded = Geo(jsonData: data)

[philprime]

Just want to point out that Decodable is considered to be decoder-agnostic, so adding a convenience method to the type would not be agnostic anymore. It also takes away the option to configure JSONDecoder in unit tests etc.

[brustolin]

It is just a convenience initializer, just like the decodeFromJSONData that it would replace, that also does not give the option to configure JSONDecoder