Searches with non-ASCII characters like umlauts or accented characters are always case-sensitive

[Archaeopteryx]

Searches with non-ASCII characters like umlauts or accented characters are always case-sensitive, e.g.

http://transvision.mozfr.org/?recherche=%C3%BCberdeckende&repo=gaia&sourcelocale=en-US&locale=de&search_type=strings_entities returns nothing

http://transvision.mozfr.org/?recherche=%C3%9Cberdeckende&repo=gaia&sourcelocale=en-US&locale=de&search_type=strings_entities finds the strings.

[pascalchevrel]

I looked into it and one suggested solution on stackoverflow is to normalize the input and data through the Transliterator extension. I am leaving the pointers here for reference:
http://stackoverflow.com/questions/14072333/php-preg-grep-and-umlaut-accent
http://php.net/class.transliterator

Normalizing the input may be a solution but would need some big changes to how we store and retrieve data on all views, so I don't think I can provide a patch in the short term.

[flodolo]

I'm trying to wrap my head around this. It definitely works by adding /u to regexps, but then str_replace fails to highlight the results, which probably means we're mixing up charsets.

Working on this for a bit.