Summary

Named HTML character references containing digits (e.g. ¹, ½, ░, ∴,  ) are not processed correctly in Angular templates. There are 24 such entities defined in the HTML spec, and none of them render properly.

Root cause

The isNamedEntityEnd function in packages/compiler/src/ml_parser/lexer.ts uses !chars.isAsciiLetter(code) to determine when the entity name ends. Since digits are not ASCII letters, the lexer stops reading the entity name prematurely when it encounters a digit character.

For example, when parsing ¹:

1. The lexer reads s, u, p (all ASCII letters ✓)
2. It hits 1 — isAsciiLetter('1') is false, so it stops
3. The captured name is sup instead of sup1
4. The next character is 1, not ;, so the entity parse fails
5. The lexer falls back to emitting & as literal text

Fix

Added !chars.isDigit(code) to the isNamedEntityEnd predicate so that digits are treated as valid continuation characters in named entity names. This matches the HTML spec, which allows alphanumeric characters in named character references.

Test plan

• Added test case 'should parse named entities containing digits' covering 6 different entities: ¹, ½, ¾, ░, ∴,  
• These cover entities with digits at the end, in the middle, and the special there4 case
• Verified all entities are already present in the NAMED_ENTITIES table in entities.ts

Fixes #51323