chore(eslint-plugin): require underscore-delimited groups in numeric literals by kriskowal · Pull Request #3263 · endojs/endo

kriskowal · 2026-05-15T01:39:33Z

Description

Adds eslint-plugin-unicorn's numeric-separators-style rule to the @endo/internal ESLint preset, enforcing underscore-delimited groupings on numeric literals. Decimal numbers of five or more digits group every three digits (1_234_567); hexadecimal, binary, and octal literals group by the rule's conventional widths (hex in twos, binary and octal in fours). The rule is autofixable, so eslint --fix applied the migration across the codebase in a separate commit. A final commit reflows seven autofix-touched files that Prettier wanted to re-wrap once the separators pushed a few expressions past the print-width threshold.

Security Considerations

None. Numeric separators are a syntactic transformation that does not change the runtime value of any literal.

Scaling Considerations

None.

Documentation Considerations

The @endo/eslint-plugin changeset notes that consumers of plugin:@endo/internal will see lint errors on numeric literals that violate the rule and must add eslint-plugin-unicorn to their devDependencies to satisfy the new peer dependency. eslint --fix rewrites violations automatically, so the migration cost downstream is one autofix run per consumer.

Testing Considerations

yarn lint is the test surface. The pre-PR run on this branch reports zero ESLint errors and zero Prettier warnings on the files this migration touches. Two pre-existing ESLint warnings on master remain unaddressed (in evasive-transform/src/index.js and ses/src/compartment.js); they predate this change and are out of scope here.

Compatibility Considerations

The rule strips separators from decimal literals below the minimum-digits threshold (1_000 becomes 1000). This matches the rule's defaults and is consistent with the convention that separators communicate magnitude only when the number is long enough to benefit from grouping.

Upgrade Considerations

None for production systems. Downstream consumers of @endo/eslint-plugin will need to install eslint-plugin-unicorn@^56.0.1 (a new peer dependency of the plugin); the published changeset describes the requirement.

changeset-bot · 2026-05-15T01:39:40Z

🦋 Changeset detected

Latest commit: 44f4d1b

The changes in this PR will be included in the next version bump.

This PR includes changesets to release 1 package

Name	Type
@endo/eslint-plugin	Minor

Not sure what this means? Click here to learn what changesets are.

Click here if you're a maintainer who wants to add another changeset to this PR

turadg

I'm not opposed to the rule but I think the config needs adjusting. Consider omitting the rule for test files and removing the hex requirement

turadg · 2026-05-15T14:22:35Z


  if (nonce === undefined) {
-    nonce = crypto.randomInt(0xffff_ffff);
+    nonce = crypto.randomInt(0xff_ff_ff_ff);


I find the pairs harder to read that quadruplets. Consider groups of four.

turadg · 2026-05-15T14:24:12Z

  if (code < 97) return code === 95;
  if (code < 123) return true;
-  if (code <= 0xffff)
+  if (code <= 0xff_ff)


Another case in point of four-segment reading better

turadg · 2026-05-15T14:24:53Z

  withInterrupt(async ({ cancelled }) => {
    await null;
-    const logCheckIntervalMs = ping !== undefined ? Number(ping) : 5_000;
+    const logCheckIntervalMs = ping !== undefined ? Number(ping) : 5000;


Why prefer no underscore for thousands?

turadg · 2026-05-15T14:26:09Z

  t.is(
    parseBigint('123456789012345678901234567890'),
-    123456789012345678901234567890n,
+    123_456_789_012_345_678_901_234_567_890n,


This is worse. The reader wants to verify the numeral sequence matches the above. Consider suppressing the rule on this file

turadg · 2026-05-15T14:27:28Z

 // Deterministic PRNG, same seed shape as other Endo fuzz tests.
-const defaultSeed = [0xb0b5c0ff, 0xeefacade, 0xb0b5c0ff, 0xeefacade];
+const defaultSeed = [
+  0xb0_b5_c0_ff, 0xee_fa_ca_de, 0xb0_b5_c0_ff, 0xee_fa_ca_de,


These visually break the mnemonic. Please suppress this line or file

turadg · 2026-05-15T14:27:42Z

 // Deterministic PRNG, same seed shape as other Endo fuzz tests.
-const defaultSeed = [0xb0b5c0ff, 0xeefacade, 0xb0b5c0ff, 0xeefacade];
+const defaultSeed = [
+  0xb0_b5_c0_ff, 0xee_fa_ca_de, 0xb0_b5_c0_ff, 0xee_fa_ca_de,


Ditto re mnemonic

turadg · 2026-05-15T14:28:44Z

  let bits = BigInt(`0x${getSuffix(encoded, skip + 1)}`);
  if (encoded.charAt(skip + 1) < '8') {
-    bits ^= 0xffffffffffffffffn;
+    bits ^= 0xff_ff_ff_ff_ff_ff_ff_ffn;


What's salient here is the length and this is harder to count than groups of four.

turadg · 2026-05-15T14:29:20Z

-  [9007199254740993n, { '@qclass': 'bigint', digits: '9007199254740993' }],
+  [9_007_199_254_740_993n, { '@qclass': 'bigint', digits: '9007199254740993' }],



Ditto the comparison point

turadg · 2026-05-15T14:30:10Z

 // See https://github.com/ocapn/ocapn/blob/main/draft-specifications/Model.md#string
-const INVALID_STRING_CHARS_START = 0xd800;
-const INVALID_STRING_CHARS_END = 0xdfff;
+const INVALID_STRING_CHARS_START = 0xd8_00;


Another case for fours: the comment above is in fours. The literal should match

kriscendobot · 2026-05-29T02:44:43Z

Mirror of endojs/endo-but-for-bots#244 (head 512438a).

…s/endo#3263 (Shape 2; numeric-separators; CONFLICTING)

… recompute, CONFLICTING→MERGEABLE); message: boatman → steward upstream cross-link

…endo#3263 (force-push; CONFLICTING->MERGEABLE; numeric-separators)

…o#3263 (clear CONFLICTING; numeric-separators)

…o#3263 (base-freshen; content unchanged)

…d bots#244 (content unchanged; MERGEABLE)

kriskowal · 2026-06-05T04:48:32Z

+      {
+        onlyIfContainsSeparator: false,
+        number: { minimumDigits: 5, groupLength: 3 },
+        binary: { minimumDigits: 0, groupLength: 4 },
+        octal: { minimumDigits: 0, groupLength: 4 },
+        hexadecimal: { minimumDigits: 0, groupLength: 4 },
+      },


@turadg The bulk of your feedback is addressed here, and with some one-off exceptions where the rules are disabled to make other representations more obviously equivalent.

turadg · 2026-06-05T16:43:15Z

   */
-  /** @this {Compartment} */
+  /**
+   * @param {...any} args


driveby but no objection

Thank you for catching this.

…literals

kriskowal mentioned this pull request May 15, 2026

chore(eslint-plugin): require underscore-delimited groups in numeric literals endojs/endo-but-for-bots#244

Closed

kriskowal requested review from kumavis and turadg May 15, 2026 01:52

kriskowal marked this pull request as ready for review May 15, 2026 01:54

kriskowal force-pushed the kriskowal-eslint-numeric-separators-style branch from b583f92 to 512438a Compare May 15, 2026 04:33

turadg reviewed May 15, 2026

View reviewed changes

kriskowal pushed a commit to kriskowal/garden that referenced this pull request May 29, 2026

dispatch: liaison — boatman re-ferries endo-but-for-bots#244 to endoj…

2e84e95

…s/endo#3263 (Shape 2; numeric-separators; CONFLICTING)

kriskowal force-pushed the kriskowal-eslint-numeric-separators-style branch from 512438a to 46de187 Compare May 29, 2026 20:45

kriskowal pushed a commit to kriskowal/garden that referenced this pull request May 29, 2026

result: boatman shape-2 re-ferry of #244 onto endojs/endo#3263 (clean…

5fa7281

… recompute, CONFLICTING→MERGEABLE); message: boatman → steward upstream cross-link

kriskowal pushed a commit to kriskowal/garden that referenced this pull request May 29, 2026

result: liaison — boatman re-ferried endo-but-for-bots#244 to endojs/…

07c2903

…endo#3263 (force-push; CONFLICTING->MERGEABLE; numeric-separators)

kriskowal force-pushed the kriskowal-eslint-numeric-separators-style branch from 46de187 to 4d039c3 Compare May 29, 2026 20:54

kriskowal enabled auto-merge May 29, 2026 20:55

kriskowal pushed a commit to kriskowal/garden that referenced this pull request Jun 3, 2026

dispatch: liaison — boatman re-ferries rebased bots#244 to endojs/end…

66c71d3

…o#3263 (clear CONFLICTING; numeric-separators)

kriskowal force-pushed the kriskowal-eslint-numeric-separators-style branch from 4d039c3 to eef8f2f Compare June 3, 2026 05:59

kriskowal requested a review from turadg June 4, 2026 05:21

kriskowal disabled auto-merge June 4, 2026 05:22

kriskowal pushed a commit to kriskowal/garden that referenced this pull request Jun 4, 2026

dispatch: liaison — boatman re-ferries rebased bots#244 to endojs/end…

2cc0808

…o#3263 (base-freshen; content unchanged)

kriskowal force-pushed the kriskowal-eslint-numeric-separators-style branch from eef8f2f to 0e861ff Compare June 4, 2026 05:35

kriskowal pushed a commit to kriskowal/garden that referenced this pull request Jun 4, 2026

result: liaison — boatman base-freshened endojs/endo#3263 from rebase…

a7bbb61

…d bots#244 (content unchanged; MERGEABLE)

kriskowal commented Jun 5, 2026

View reviewed changes

turadg approved these changes Jun 5, 2026

View reviewed changes

kriskowal added 4 commits June 5, 2026 10:19

chore(eslint-plugin): require underscore-delimited groups in numeric …

c423ed3

…literals

fix(ses): Deduplicate dist types

d45818c

fix(ses): Consolidate Compartment jsdoc comments

5065e72

chore: Update yarn.lock

44f4d1b

kriskowal force-pushed the kriskowal-eslint-numeric-separators-style branch from 0e861ff to 44f4d1b Compare June 6, 2026 04:01

kriskowal enabled auto-merge June 6, 2026 04:02

kriskowal merged commit 5865ff1 into master Jun 6, 2026
16 of 17 checks passed

kriskowal deleted the kriskowal-eslint-numeric-separators-style branch June 6, 2026 04:05

kriskowal mentioned this pull request Jun 6, 2026

Version Packages #3290

Open

		[9007199254740993n, { '@qclass': 'bigint', digits: '9007199254740993' }],
		[9_007_199_254_740_993n, { '@qclass': 'bigint', digits: '9007199254740993' }],

Conversation

kriskowal commented May 15, 2026

Description

Security Considerations

Scaling Considerations

Documentation Considerations

Testing Considerations

Compatibility Considerations

Upgrade Considerations

Uh oh!

changeset-bot Bot commented May 15, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

🦋 Changeset detected

Uh oh!

turadg left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

kriscendobot commented May 29, 2026

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

changeset-bot Bot commented May 15, 2026 •

edited

Loading