mb_ereg functionality is provided via Oniguruma RegEx library and not via PCRE. mb_regex_encoding() does only support a subset of encoding names, compared to mb_list_encodings() and mb_encoding_aliases().Currently the following names are supported (case-insensitive):UCS-4UCS-4LEUTF-32UTF-32BEUTF-32LEUTF-16UTF-16BEUTF-16LEUTF-8utf8ASCIIUS-ASCIIEUC-JPeucJPx-euc-jpSJISeucJP-winSJIS-winCP932MS932Windows-31JISO-8859-1ISO-8859-2ISO-8859-3ISO-8859-4ISO-8859-5ISO-8859-6ISO-8859-7ISO-8859-8ISO-8859-9ISO-8859-10ISO-8859-13ISO-8859-14ISO-8859-15ISO-8859-16EUC-CNEUC_CNeucCNgb2312EUC-TWEUC_TWeucTWBIG-5CN-BIG5BIG-FIVEBIGFIVEEUC-KREUC_KReucKRKOI8-RKOI8RThe list is a mixture of base names and aliases and applies to PHP 5.4.45 (Oniguruma lib v4.7.1), PHP 5.6.31 (v5.9.5), PHP 7.0.22 (v5.9.6) and PHP 7.1.8 (v5.9.6). Be aware of the inconsistency: mb_regex_encoding() accepts for example the base name 'UTF-8' and its only alias 'utf8', but it does not accept aliases 'utf16', 'utf32' or 'latin1'.Additionally note, that the informal name/alias 'latin9' for ISO/IEC 8859-15:1999 (including the Euro sign on 0xA4) is also not known by mb_list_encodings(). It can only be adressed as 'ISO-8859-15' or 'ISO_8859-15' and for mb_regex_encoding() solely as 'ISO-8859-15'.