[TYPO3-core] Re: FYI72: #15130: Backend search does not find non–latin1 characters

[Martin Kutschker]

Am 15.07.2010 11:07, schrieb Dmitry Dulepov:
> Hi!
> 
> This is SVN patch request.
> 
> Type: bug
> 
> BT reference: http://bugs.typo3.org/view.php?id=15130
> 
> Branches: 4.3, 4.4, trunk
> 
> Problem: If the user types non–latin characters to the Backend search,
> nothing will be found. I used latvian (ā, ē, ī, š, č) characters in
> combination with regular english characters. Also russian characters cannot
> be found. This becomes a real problem if you have thousands of users with
> russian names and want to find them by name in the Backend. The problem is
> in the incorrect JSON creation for search.
> 
> Solution: remove unnecessary rawurlencode() on JSON parameters. That
> creates double encoding for any non–latin1 data.

I hope that just removing the call is enough. But as I have written repeatedly: rawurlencode() is
NOT to be used on non-ASCII data. Anyway, you will need it or its's "newer" cousin decodeUrl only
when dealing with URLs (and GET data). Don't use it for POST or JS strings.

Note to the clean up team: remove all the unnecessary calls to rawurlencode in the BE and get rid of
the JS string (de-)composition with charCodeAt().

Masi