AOT: mb_strwidth()/mb_strimwidth() refuse runtime encoding — LogicException (leftover of #34264, ext/mbstring/mbstring.c)

[PurHur]

Category

bug · AOT · stdlib · mbstring · NestedJIT · php-src-strict

Problem

After #34858 / #34875 / #34880 wave, mb_strwidth() / mb_strimwidth() still refuse a runtime encoding string (JitMbStrwidth throws LogicException: … encoding must be a string literal). Invalid encoding literals also abort compile via VmMbstring::strwidth fold instead of a catchable runtime ValueError.

Repro	Zend 8.2+	AOT (master @ 8a29eaf)
mb_strwidth('über', enc()) with enc()=>'UTF-8'	int(4)	compile: LogicException encoding must be a string literal
mb_strimwidth('übercafé', 0, 5, '...', enc())	üb...	same LogicException
$e='NOPE'; try { mb_strwidth('x',$e); } catch (ValueError)	catchable ValueError	compile-time uncaught ValueError

./script/aot-smoke.sh is 8/8. Peer closed: #34858, #34875, #34880.

php-src reference

• php/php-src ext/mbstring/mbstring.c — PHP_FUNCTION(mb_strwidth) / PHP_FUNCTION(mb_strimwidth)

PHP implementation target

• ext/mbstring/JitMbStrwidth.php — NestedJIT encoding ptr + assertEncodingArgv (peer JitMbStrcut / AOT: mb_substr()/mb_strcut() refuse runtime encoding — LogicException (leftover of #34256, ext/mbstring/mbstring.c) #34875); only fold when encoding is a supported canon
• ext/mbstring/MbStrwidthJitHelper.php — int-returning assertEncodingArgv (inline canon; no private helper); Argument Php stdlib specs implementation #2 for strwidth, Stdlib: floatval, boolval, gettype(null) and expanded PHPT specs #5 for strimwidth
• lib/JIT/Builtin/MbStrwidth.php — register assert helper
• No new runtime/*.c

Repro

./script/docker-exec.sh -- bash -lc 'source script/php-env.sh
export PHP_COMPILER_HELPER_RUNTIME_O=0
php bin/compile.php -o /tmp/t.bin test/repro/mb_strwidth_runtime_encoding_aot.php && /tmp/t.bin
php test/repro/mb_strwidth_runtime_encoding_aot.php
'

Done when

• AOT matches Zend for runtime UTF-8 encoding on strwidth/strimwidth
• Invalid encoding is catchable ValueError at runtime (not compile abort)
• Targeted phpunit green; ./script/aot-smoke.sh stays 8/8
• No new runtime/*.c

[PurHur]

claim: stdlib-aot agent — starting this run (#32122/#31968 children/#34881 closed; aot-smoke 8/8; NestedJIT runtime encoding for mb_strwidth/mb_strimwidth leftover of #34264).

[PurHur]

claim: implementer-worker-lane-A — starting this run (lane %3==0; #32122/#31968/#34881 closed; aot-smoke 8/8). Implementing mb_strwidth/mb_strimwidth NestedJIT assertEncodingArgv (peer #34875/#34880).