💥 remove Object.* Polyfills

[thomas-lebeau]

Motivation

Removed unnecessary polyfills for v6, resulting in a slight improvement in the bundle sizes.

Changes

• Remove objectValues and objectEntries polyfills. However I'm keeping the wrapper method because their mangling result in smaller bundle size than using Object.values() every time it is used.
• Remove Object.assign polyfills in favor of object spread operator.

Testing

• Local
• Staging
• Unit
• End to end

---

I have gone over the contributing documentation.

[codecov-commenter]

Codecov Report

Attention: Patch coverage is 91.13924% with 7 lines in your changes missing coverage. Please review.

Project coverage is 93.76%. Comparing base (c94a365) to head (a0617ef).

Files	Patch %	Lines
packages/core/src/boot/init.ts	33.33%	2 Missing ⚠️
packages/logs/src/boot/logsPublicApi.ts	33.33%	2 Missing ⚠️
packages/rum-core/src/boot/rumPublicApi.ts	33.33%	2 Missing ⚠️
.../resource/retrieveInitialDocumentResourceTiming.ts	66.66%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##               v6    #2908      +/-   ##
==========================================
- Coverage   93.76%   93.76%   -0.01%     
==========================================
  Files         267      267              
  Lines        7378     7372       -6     
  Branches     1659     1658       -1     
==========================================
- Hits         6918     6912       -6     
  Misses        460      460              

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[cit-pr-commenter]

Bundles Sizes Evolution

📦 Bundle Name	Base Size	Local Size	𝚫	𝚫%	Status
Rum	149.42 KiB	149.35 KiB	-71 B	-0.05%	✅
Logs	50.78 KiB	50.66 KiB	-122 B	-0.23%	✅
Rum Slim	99.38 KiB	99.31 KiB	-67 B	-0.07%	✅
Worker	25.16 KiB	25.16 KiB	0 B	0.00%	✅

🚀 CPU Performance

Action Name	Base Average Cpu Time (ms)	Local Average Cpu Time (ms)	𝚫
addglobalcontext	0.002	0.002	0.001
addaction	0.031	0.037	0.006
adderror	0.030	0.034	0.004
addtiming	0.001	0.001	0.000
startview	0.847	1.003	0.156
startstopsessionreplayrecording	0.824	0.727	-0.097
logmessage	0.020	0.018	-0.002

🧠 Memory Performance

Action Name	Base Consumption Memory (bytes)	Local Consumption Memory (bytes)	𝚫 (bytes)
addglobalcontext	22.30 KiB	21.85 KiB	-461 B
addaction	71.21 KiB	71.64 KiB	435 B
adderror	85.84 KiB	85.09 KiB	-769 B
addtiming	19.49 KiB	19.08 KiB	-415 B
startview	336.36 KiB	337.16 KiB	826 B
startstopsessionreplayrecording	16.13 KiB	16.00 KiB	-133 B
logmessage	71.88 KiB	71.57 KiB	-322 B

🔗 RealWorld

[thomas-lebeau]

looks weird but it's because this some type is inferred as ReplicaUserConfiguration in another module (remoteConfiguration.ts) but typescript can't name it.

@datadog/browser-rum-core: src/domain/configuration/remoteConfiguration.ts(22,17): error TS4058: Return type of exported function has or is using name 'ReplicaUserConfiguration' from external module "/Users/thomas.lebeau/go/src/github.com/DataDog/browser-sdk/packages/core/cjs/domain/configuration/configuration" but cannot be named.

[nulrich]

assign({}, initConfiguration, remoteInitConfiguration) return type is the intersection of RumInitConfiguration and RumRemoteConfiguration. So ReplicaUserConfiguration is not exposed.

And I think that { ...initConfiguration, ...remoteInitConfiguration } export a new type containing all the props from both object types, including non exported ReplicaUserConfiguration.

Maybe that's the reason.

[thomas-lebeau]

Somehow, if T is undefined, then TS will infer {...object} as undefined, insead of {}

see TS playground

[thomas-lebeau]

type is unnecessary here, actually TS errors with: 'type' is specified more than once, so this usage will be overwritten. ts(2783)

spreading clickActionBase will always overwrites type

[nulrich]

assign({}, initConfiguration, remoteInitConfiguration) return type is the intersection of RumInitConfiguration and RumRemoteConfiguration. So ReplicaUserConfiguration is not exposed.

And I think that { ...initConfiguration, ...remoteInitConfiguration } export a new type containing all the props from both object types, including non exported ReplicaUserConfiguration.

Maybe that's the reason.