Do the wikitext: {{#expr:ģ}}
Expected behaviour:
Expression error: Unrecognized punctuation character "ģ".
Actual behaviour:
Expression error: Unrecognized punctuation character "�".
Maybe something is converting from iso-8859-1?
Version: unspecified
Severity: normal
| Subject | Repo | Branch | Lines +/- | |
|---|---|---|---|---|
| Enable non-ascii letters in expression error | mediawiki/extensions/ParserFunctions | master | +2 -1 |
Is there any problem with Validator:cleanup() function in the core which converts non ascii character normal to C?
No.
Its because on line 200 of Expr.php, $char = $expr[$p];, the expression parser walks through the expression a byte at a time. In the case of a multibyte unicode character (including 'LATIN SMALL LETTER G WITH CEDILLA' (U+0123)), this will only look at the first byte, so the error message includes only the byte 0xC4 which is an invalid unicode code sequence, so the Validator::cleanUp() method correctly converts this to a replacement sequence.
One possible fix would be on line 336 to replace Validator::cleanUp( $char ) with Validator::cleanUp( mb_substr( $expr, $p, 1 ) )
thankyou @Bawolff presently it is executing correctly in my localhost can I make a patch of it in Gerrit?
Change 407596 had a related patch set uploaded (by Rammanojpotla; owner: Rammanoj):
[mediawiki/extensions/ParserFunctions@master] Enable non-ascii letters in expression error
Change 407596 merged by jenkins-bot:
[mediawiki/extensions/ParserFunctions@master] Enable non-ascii letters in expression error