Update recordid.ts

[CLoaKY233]

Fix: Proper Escaping of Long Numeric String Record IDs

Issue Description

ISSUE ---> #389 (comment)
When handling RecordId instances with long numeric strings as identifiers (15 or more digits), the current implementation doesn't properly escape these values. This causes inconsistency with SurrealDB's native behavior, which wraps such identifiers in angle brackets (⟨⟩).

Current Behavior

new RecordId('foo', '125813199042576601589342522460260755').toString()
// Returns: foo:125813199042576601589342522460260755

Expected Behavior

new RecordId('foo', '125813199042576601589342522460260755').toString()
// Should return: foo:⟨125813199042576601589342522460260755⟩

Changes Made

1. Modified escapeIdPart function to detect and properly escape long numeric string IDs
2. Updated StringRecordId constructor to handle long numeric string IDs consistently
3. Added specific handling for IDs matching the pattern /^\d{15,}$/

Technical Details

• Added check for numeric strings of 15 or more digits
• Implemented wrapping with angle brackets (⟨⟩) for matching IDs
• Preserved existing behavior for all other ID types
• Maintains backward compatibility with existing implementations

Testing

Tested with various ID types including:

• Regular string IDs
• Short numeric strings
• Long numeric strings
• UUIDs
• Complex objects and arrays

Examples

// Long numeric string ID
new RecordId('table', '125813199042576601589342522460260755')
// => table:⟨125813199042576601589342522460260755⟩

// Regular string ID (unchanged behavior)
new RecordId('table', 'regular-id')
// => table:regular-id

// Numeric ID (unchanged behavior)
new RecordId('table', 123)
// => table:123

[kearfy]

Thank you for your PR @CLoaKY233! Left two comments 🙏

[CLoaKY233]

Following the review feedback, I've simplified the fix to focus on the root cause:

1. Changed isOnlyNumbers function to use regex instead of parseInt:

function isOnlyNumbers(str: string): boolean {
    return /^\d+$/.test(str.replace(/_/g, ''));
}

Key improvements:

• Replaced parseInt-based check with regex pattern /^\d+$/
• Added global flag to underscore replacement (/_/g) to handle multiple underscores
• Removed unnecessary string conversion and comparison
• More reliable detection of numeric strings of any length

[kearfy]

Awesome @CLoaKY233, thank you so much! One more favor, could you run bun quality:apply:unsafe to fix the "Code Quality" test? Thank you 🙏

[CLoaKY233]

Done Sir! :D

[kearfy]

Thanks for bearing with me @CLoaKY233! Merged it!