perf: optimize splitCookiesString with charCode checks

[benfavre]

Summary

Optimize splitCookiesString in packages/next/src/server/web/utils.ts by replacing per-character regex matching with integer charCode comparisons.

Problem

The skipWhitespace() helper creates a /\s/ regex on every call and uses charAt() which allocates a single-character string. This function is called in a tight character-scanning loop for every cookie header on every request.

Changes

• Replace /\s/.test(cookiesString.charAt(pos)) with isWhitespace(cookiesString.charCodeAt(pos)) using a simple integer comparison helper
• Replace charAt() comparisons in notSpecialChar() and the main loop with charCodeAt() integer comparisons
• Zero behavioral change — same characters are matched

Why this matters

Inspired by TanStack Start's 5.5x SSR throughput: hot-path micro-optimizations compound. Regex creation and string allocation in tight loops are exactly the kind of overhead that adds up across thousands of requests.

Test plan

• Verify cookie parsing works correctly with compound set-cookie headers
• Verify middleware cookie forwarding still works
• No regressions in cookie handling

🤖 Generated with Claude Code

[nextjs-bot]

Allow CI Workflow Run

• approve CI run for commit: a60fcc9

Note: this should only be enabled once the PR is ready to go and can only be enabled by a maintainer

[benfavre]

Test Verification

• utils.test.ts: 4/4 passed
• test/unit/split-cookies-string.test.ts: 10/10 passed
• Vertical tab (charCode 11) added per VADE feedback

All tests run on the perf/combined-all branch against canary. Total: 203 tests across 13 suites, all passing.