chore: use stable empty references

[n3ps]

Description

When selectors return empty arrays or objects using the ?? or || operators, they create new instances every time the function is called if the value is undefined/null. This breaks reference equality and memoization and can cause excess re-renders.

Changelog

CHANGELOG entry: chore: use stable empty references

Related issues

Fixes:

Manual testing steps

1. Go to this page...

Screenshots/Recordings

Before

useSelector(state => state.some.collection || []) // new array each time

After

// stable instance
const EMPTY = []

useSelector(state => state.some.collection || EMPTY)

Pre-merge author checklist

• I've followed MetaMask Contributor Docs and MetaMask Extension Coding Standards.
• I've completed the PR template to the best of my ability
• I’ve included tests if applicable
• I’ve documented my code using JSDoc format if applicable
• I’ve applied the right labels on the PR (see labeling guidelines). Not required for external contributors.

Pre-merge reviewer checklist

• I've manually tested the PR (e.g. pull and build branch, run the app, test code being changed).
• I confirm that this PR addresses all acceptance criteria described in the ticket it closes and includes the necessary testing evidence such as recordings and or screenshots.

---

Note

Replaces ad-hoc []/{} fallbacks with shared frozen constants across selectors to stabilize memoization and references.

• Selectors:
  • Alerts: selectAlerts now returns EMPTY_ARRAY fallback.
  • Assets: use EMPTY_OBJECT for getMetamaskState and in balance-related selectors (account tree, rates, tokens, currency rates) instead of {}; remove local EMPTY_OBJ in favor of shared constant.
  • NFTs: default allNftContracts/allNfts to EMPTY_OBJECT.
  • General selectors: selectNftsByChainId and selectNonZeroUnusedApprovalsAllowList now return EMPTY_ARRAY when empty.
• Shared utils:
  • Add ui/selectors/shared.ts exporting frozen EMPTY_ARRAY and EMPTY_OBJECT; import and use where needed.

^{Written by Cursor Bugbot for commit 9272811. This will update automatically on new commits. Configure here.}

[github-actions]

CLA Signature Action: All authors have signed the CLA. You may need to manually re-run the blocking PR check if it doesn't pass in a few minutes.

[metamaskbot]

📊 Page Load Benchmark Results

Current Commit: 970ded4 | Date: 10/16/2025

📄 Localhost MetaMask Test Dapp

Samples: 100

Summary

• pageLoadTime-> current mean value: 1.06s (±77ms) 🟡 | historical mean value: 1.05s ⬆️ (historical data)
• domContentLoaded-> current mean value: 747ms (±74ms) 🟢 | historical mean value: 734ms ⬆️ (historical data)
• firstContentfulPaint-> current mean value: 77ms (±14ms) 🟢 | historical mean value: 80ms ⬇️ (historical data)

📈 Detailed Results

Metric	Mean	Std Dev	Min	Max	P95	P99
pageLoadTime	1.06s	77ms	1.02s	1.41s	1.30s	1.41s
domContentLoaded	747ms	74ms	711ms	1.08s	959ms	1.08s
firstPaint	77ms	14ms	60ms	200ms	88ms	200ms
firstContentfulPaint	77ms	14ms	60ms	200ms	88ms	200ms
largestContentfulPaint	0ms	0ms	0ms	0ms	0ms	0ms

---

Results generated automatically by MetaMask CI

[metamaskbot]

Builds ready [970ded4]

• builds: chrome, firefox
• builds (beta): chrome, firefox
• builds (flask): chrome, firefox
• builds (test): chrome, firefox
• builds (test-flask): chrome, firefox
• build viz: Build System
• bundle size: Bundle Size Stats
• user-actions-benchmark: User Actions Stats
• storybook: Storybook
• typescript migration: Dashboard
• all artifacts
• bundle viz:

  • background: 0, 1, 2, 3, 4, 5, 6, 7, 8
  • common: 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13
  • ui: 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14
  • content-script: 0
  • offscreen: 0

UI Startup Metrics (1250 ± 72 ms)

Platform	BuildType	Page	Metric	Mean (ms)	Min (ms)	Max (ms)	Std Dev (ms)	P 75 (ms)	P 95 (ms)
Chrome	Browserify	Home	uiStartup	1250	1110	1514	72	1291	1388
			load	1074	956	1272	68	1116	1212
			domContentLoaded	1068	952	1266	68	1108	1206
			domInteractive	19	13	123	12	17	39
			firstPaint	751	80	1284	415	1093	1162
			backgroundConnect	251	237	290	9	255	272
			firstReactRender	26	18	48	7	27	43
			getState	15	5	39	7	19	32
			initialActions	5	1	49	6	6	15
			loadScripts	822	693	1001	66	859	944
			setupStore	9	6	22	3	10	15
	Webpack	Home	uiStartup	834	712	1155	72	852	1015
			load	628	577	969	71	631	829
			domContentLoaded	619	571	942	67	625	782
			domInteractive	16	11	66	10	14	37
			firstPaint	173	54	928	171	190	604
			backgroundConnect	22	10	67	9	27	42
			firstReactRender	27	16	54	8	31	35
			getState	10	4	38	5	11	15
			initialActions	3	0	13	2	4	6
			loadScripts	617	569	931	65	623	770
			setupStore	9	5	16	3	11	13
Firefox	Browserify	Home	uiStartup	1456	1275	1919	118	1517	1663
			load	1221	1073	1440	82	1276	1377
			domContentLoaded	1221	1073	1440	82	1276	1376
			domInteractive	106	26	292	50	114	247
			firstPaint	NaN	NaN	NaN	NaN	NaN	NaN
			backgroundConnect	33	21	68	10	38	57
			firstReactRender	30	25	64	6	31	40
			getState	12	4	206	27	8	18
			initialActions	5	1	64	7	4	15
			loadScripts	1199	1053	1411	81	1250	1340
			setupStore	12	5	65	11	11	46
	Webpack	Home	uiStartup	1602	1380	2108	139	1651	1945
			load	1356	1207	1607	87	1404	1512
			domContentLoaded	1355	1206	1607	87	1404	1512
			domInteractive	100	30	389	72	97	344
			firstPaint	NaN	NaN	NaN	NaN	NaN	NaN
			backgroundConnect	37	19	92	13	42	60
			firstReactRender	35	26	98	11	36	53
			getState	10	4	56	9	9	30
			initialActions	12	1	228	37	4	125
			loadScripts	1332	1180	1576	86	1377	1492
			setupStore	16	6	193	25	11	61

Bundle size diffs [🚨 Warning! Bundle size has increased!]

• background: 58 Bytes (0%)
• ui: 329 Bytes (0%)
• common: 662 Bytes (0.01%)

[MajorLift]

When a variable is destructured from ?? {}, the purpose is to safely assign it with undefined without triggering the "property of undefined" JS TypeError.

Since the reference of the empty object in these cases are never used beyond each destructuring assignment, and also not by React, these aren't cases where the EMPTY_OBJECT substitution would be useful.

[MajorLift]

Intermediate variables not used outside of the selector, or downstream by React.

[MajorLift]

createSelector memoizes its output based on its input arguments.

https://reselect.js.org/api/weakmapmemoize/

So the references of create{,DeepEqual}Selector output are already stable without needing these changes.

[MajorLift]

👍 Similar optimizations might be more impactful if they're made in React components and hooks to reactive variables that require referential equality rather than upstream in the selectors.

[n3ps]

createSelector memoizes its output based on its input arguments.

https://reselect.js.org/api/weakmapmemoize/

So the references of create{,DeepEqual}Selector output are already stable without needing these changes.

Thanks for the review @MajorLift! Updated the objects, but I think the arrays are still valid:

https://reselect.js.org/usage/handling-empty-array-results/

[metamaskbot]

📊 Page Load Benchmark Results

Current Commit: 9272811 | Date: 10/16/2025

📄 Localhost MetaMask Test Dapp

Samples: 100

Summary

• pageLoadTime-> current mean value: 1.00s (±74ms) 🟡 | historical mean value: 1.05s ⬇️ (historical data)
• domContentLoaded-> current mean value: 702ms (±85ms) 🟢 | historical mean value: 737ms ⬇️ (historical data)
• firstContentfulPaint-> current mean value: 79ms (±40ms) 🟢 | historical mean value: 77ms ⬆️ (historical data)

📈 Detailed Results

Metric	Mean	Std Dev	Min	Max	P95	P99
pageLoadTime	1.00s	74ms	956ms	1.32s	1.24s	1.32s
domContentLoaded	702ms	85ms	659ms	1.24s	944ms	1.24s
firstPaint	79ms	40ms	60ms	472ms	88ms	472ms
firstContentfulPaint	79ms	40ms	60ms	472ms	88ms	472ms
largestContentfulPaint	0ms	0ms	0ms	0ms	0ms	0ms

---

Results generated automatically by MetaMask CI

[metamaskbot]

Builds ready [9272811]

• builds: chrome, firefox
• builds (beta): chrome, firefox
• builds (flask): chrome, firefox
• builds (test): chrome, firefox
• builds (test-flask): chrome, firefox
• build viz: Build System
• bundle size: Bundle Size Stats
• user-actions-benchmark: User Actions Stats
• storybook: Storybook
• typescript migration: Dashboard
• all artifacts
• bundle viz:

  • background: 0, 1, 2, 3, 4, 5, 6, 7, 8
  • common: 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13
  • ui: 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14
  • content-script: 0
  • offscreen: 0

UI Startup Metrics (1260 ± 57 ms)

Platform	BuildType	Page	Metric	Mean (ms)	Min (ms)	Max (ms)	Std Dev (ms)	P 75 (ms)	P 95 (ms)
Chrome	Browserify	Home	uiStartup	1260	1146	1464	57	1293	1347
			load	1082	979	1224	51	1113	1170
			domContentLoaded	1075	951	1218	52	1107	1163
			domInteractive	18	14	45	5	17	26
			firstPaint	630	80	1231	444	1102	1154
			backgroundConnect	258	242	372	13	262	269
			firstReactRender	28	17	129	14	29	42
			getState	16	6	38	8	21	31
			initialActions	6	0	57	7	7	19
			loadScripts	824	710	970	50	854	915
			setupStore	10	7	45	5	11	19
	Webpack	Home	uiStartup	807	677	1017	65	836	931
			load	610	545	847	73	616	780
			domContentLoaded	602	540	842	72	607	774
			domInteractive	16	11	70	10	14	37
			firstPaint	172	51	781	163	167	569
			backgroundConnect	21	10	41	7	27	34
			firstReactRender	25	16	36	7	32	35
			getState	9	4	17	3	11	15
			initialActions	3	0	8	2	4	8
			loadScripts	599	538	840	72	605	772
			setupStore	10	5	16	3	12	15
Firefox	Browserify	Home	uiStartup	1427	1229	2034	127	1488	1665
			load	1206	1059	1453	82	1269	1347
			domContentLoaded	1206	1059	1453	82	1269	1347
			domInteractive	101	34	288	47	107	219
			firstPaint	NaN	NaN	NaN	NaN	NaN	NaN
			backgroundConnect	33	22	90	10	38	51
			firstReactRender	30	24	59	6	31	44
			getState	9	5	75	10	8	25
			initialActions	4	2	19	3	4	12
			loadScripts	1185	1044	1427	81	1246	1329
			setupStore	17	6	235	34	10	59
	Webpack	Home	uiStartup	1622	1440	2013	120	1681	1894
			load	1383	1231	1630	89	1438	1563
			domContentLoaded	1383	1230	1630	89	1437	1562
			domInteractive	101	31	410	69	99	333
			firstPaint	NaN	NaN	NaN	NaN	NaN	NaN
			backgroundConnect	39	21	121	18	43	76
			firstReactRender	37	26	88	14	36	85
			getState	9	4	53	5	10	16
			initialActions	6	2	58	9	4	17
			loadScripts	1357	1216	1602	86	1404	1546
			setupStore	13	5	76	12	12	39

Bundle size diffs [🚀 Bundle size reduced!]

• background: -71.86 KiB (-1.58%)
• ui: 1.38 KiB (0.02%)
• common: 57.44 KiB (0.69%)

[MajorLift]

LGTM! Going forward, React Compiler should take care of this codebase-wide for most relevant cases, but it would be good to see if there are exceptions where manually applying these changes would make a difference for performance.