Bug #52211 (iconv() returns part of string on error) [ext/iconv/tests/bug52211.phpt

Start by running ext/iconv/tests/bug52211.phpt and reading the related iconv extension test and implementation paths. Check how the test selects charsets and how the conversion error is reported; done means the test produces the expected illegal-character notice and false result on supported implementations.

Description

The following code:

// According to POSIX 2024, the to/from charset names are
// implementation-defined. To keep this test true to its original
// purpose, we retain the charsets used in bug 52211, but only when
// the implementation is known to support them. Otherwise we default
// both to ASCII, which should be supported everywhere (in particular
// on musl) yet still considers the input invalid.
$from_charset = "ASCII";
$to_charset = "ASCII";

if (ICONV_IMPL == "libiconv" || ICONV_IMPL == "glibc") {
    $from_charset = "CP850";
    $to_charset = "ISO-8859-1";
}

$str = "PATHOLOGIES MÉDICO-CHIRUR. ADUL. PL";
$str_iconv = iconv($from_charset, $to_charset, $str );
var_dump($str_iconv);
<?php

Resulted in this output:

Warning: iconv(): Wrong encoding, conversion from "CP850" to "ISO-8859-1" is not allowed in /opt/php/8.6/php-8.6.0beta3/ext/iconv/tests/bug52211.php on line 18
bool(false)

But I expected this output instead:

Notice: iconv(): Detected an illegal character in input string in %s on line %d
bool(false)

PHP Version

PHP 8.6.0beta3 (cli) (built: Sep 24 2026 01:10:42) (NTS DEBUG)
Copyright © The PHP Group and Contributors
Zend Engine v4.6.0-dev, Copyright © Zend by Perforce
    with Zend OPcache v8.6.0beta3, Copyright ©, by Zend by Perforce

Operating System

Fedora Linux 44 (Container Image)