Stdlib: html_entity_decode() default flags — named entities like   not decoded (ext/standard/html.c)

[PurHur]

Category

stdlib — php-src-strict

Problem

With default flags (ENT_COMPAT | ENT_HTML401), html_entity_decode(' ') must decode named HTML entities to their UTF-8 bytes. This compiler leaves   literal.

Verified 2026-06-23 on VM (php bin/vm.php).

php-src reference

• ext/standard/html.c — php_html_entities() / php_unescape_html_entities()
• ext/standard/html_tables.h — named entity table

Repro

test/repro/maintainer_gap_html_entity_decode_nbsp.php

php test/repro/maintainer_gap_html_entity_decode_nbsp.php           # exit 0
php bin/vm.php test/repro/maintainer_gap_html_entity_decode_nbsp.php  # exit 1

Zend PHP: PASS (\u{00A0})

This compiler:   unchanged

Implementation hint

• ext/standard/VmString.php / html_entity_decode.php — wire default-flag named-entity decode through shared HTML entity tables (PHP-in-PHP)
• Distinct from combined ENT_* flag handling (Stdlib: html_entity_decode() combined ENT_* flags — LogicException not decoded output (ext/standard/html.c) #10519)
• No new runtime/*.c entity tables

Done when

• Repro exits 0 on VM
•  , <, > decode with default flags
• Targeted compliance guard under test/compliance/cases/stdlib/
• ./script/ci-fast.sh green

Links

• Related Stdlib: html_entity_decode() combined ENT_* flags — LogicException not decoded output (ext/standard/html.c) #10519, Stdlib: htmlspecialchars_decode() — enum case operand must TypeError (ext/standard/string.c, php-src-strict) #6163
• refs Epic: Self-host critical path — compiler compiles itself (M3→M5) #1492

[PurHur]

claim: PHP Compiler STD Lib — starting this run

[PurHur]

claim released: repro green on master — html_entity_decode(' ') returns U+00A0 (c2a0); compliance html_entity_decode_nbsp.phpt passes VM.

[PurHur]

Stale: verified green on master (2026-07-12, lane-C hygiene run).

./script/docker-exec.sh -- bash -lc 'source script/php-env.sh
php bin/vm.php -r "echo bin2hex(html_entity_decode(\" \")), \"\\n\";"
'

Output: c2a0 (U+00A0) — named entity   decoded on VM.