Stdlib: grapheme_strimwidth() missing — PHP 8.4 intl width trim (ext/intl/grapheme_string.c)

[PurHur]

Category

stdlib · php-src-strict

Problem

PHP 8.4 adds grapheme_strimwidth() — grapheme-cluster-aware string trim by display width (intl). Closed umbrella #6081 tracked both grapheme_strimwidth() and mb_str_pad(); mb_str_pad() is now registered on VM (function_exists → true, verified 2026-06-18) but grapheme_strimwidth() remains absent (function_exists → false).

php-src reference

• php/php-src ext/intl/grapheme/grapheme_string.c — PHP_FUNCTION(grapheme_strimwidth)
• php/php-src ext/intl/grapheme/grapheme_string.stub.php
• Related: Stdlib: grapheme_str_split() — PHP 8.4 grapheme cluster split (ext/intl/grapheme) #5958 grapheme_str_split, Stdlib: grapheme_extract() — missing intl grapheme substring extractor (ext/intl/grapheme) #6023 grapheme_extract

Repro

<?php
var_dump(function_exists('grapheme_strimwidth'));
echo grapheme_strimwidth('こんにちは', 0, 3, '...'), "\n";

./script/docker-exec.sh -- bash -lc 'source script/php-env.sh
php bin/vm.php -r "var_dump(function_exists(\"grapheme_strimwidth\"));"
php bin/vm.php repro_grapheme_strimwidth.php
'

Engine	function_exists	Output
Zend PHP 8.4+	true	grapheme-trimmed string with ellipsis
VM today (verified 2026-06-18)	false	undefined function

Scope (this repo)

Path	Work
ext/intl/ or ext/standard/	grapheme_strimwidth.php VM handler
Width logic	reuse Utf8JitHelper / mbstring width helpers where possible
Registration	ext/intl/Module.php or ext/standard/Module.php
Tests	test/compliance/cases/stdlib/grapheme_strimwidth.phpt

PHP-in-PHP first: implement width trim in PHP; thin ICU/libc only if proven necessary.

Done when

• function_exists('grapheme_strimwidth') true on VM
• Repro output matches Zend PHP 8.4+ for UTF-8 sample
• Enum case operands → TypeError (php-src-strict)
• ./script/ci-fast.sh --filter grapheme_strimwidth green
• Update stale Stdlib: mb_str_pad() — multibyte string padding (ext/mbstring/mbstring.c parity) #4006 / Stdlib: grapheme_strimwidth() / mb_str_pad() — PHP 8.4 string helpers missing (ext/intl, ext/mbstring) #6081 notes if needed

Related

• Stdlib: grapheme_strimwidth() / mb_str_pad() — PHP 8.4 string helpers missing (ext/intl, ext/mbstring) #6081 (closed umbrella) · Stdlib: mb_str_pad() — multibyte string padding (ext/mbstring/mbstring.c parity) #4006 (mb_str_pad landed) · Epic: Self-host critical path — compiler compiles itself (M3→M5) #1492

[PurHur]

claim: cursor-agent-lane-b — starting this run

[PurHur]

claim released: grapheme_strimwidth() is not in php-src PHP 8.4 intl grapheme API (php.net ref.intl.grapheme lists no such function). Issue may need close/re-scope — switching to another lane-B item.

[PurHur]

claim: PHP Compiler STD Lib Advanced — batch (DOM managed-property fetch #14419 follow-up, #9793, #4006)

[PurHur]

Merged via #15928 — VmGrapheme::strimwidth() + grapheme_strimwidth builtin behind intl + PHP 8.4 profile gate. Unit tests green; full function_exists path when ext/intl advertises.

[PurHur]

claim: PHP Compiler STD Lib — starting this run (JIT/AOT compile-time fold for grapheme_strimwidth, #9793)

[PurHur]

Merged via #16377.

Verification (docker-exec):

• vendor/bin/phpunit --filter VmGraphemeStrimwidthTest — OK (4 tests)
• php bin/compile.php -l test/fixtures/aot/compile-only/grapheme_strimwidth_literals.php — exit 0
• VM semantics unchanged: VmGrapheme::strimwidth('hello', 0, 10) → hello

JIT/AOT now uses compile-time fold via JitGrapheme::tryStrimwidthFold() (runtime lowering deferred, same as grapheme_substr() #3352). Builtin remains gated on extension_loaded('intl') + PHP 8.4 profile per #11472.