Double Conversion of utf-8 calling ./kinosearch

Dear Maintainer and Users of Foswiki with Kinosearch,

I have the following problem with "Umlaute" under Foswiki 2.1.6. They come out scrambled in the output, links to filenames of attachments with "Umlaute" don't match (what is the more serious problem). It comes out perfect with Foswiki 1.1.9.

Distributions used: Debian 6 and Debian 10

Versions: Foswiki 1.1.9 and 2.1.6

Plugins KinosearchContrib 1.31 (last Version identical in both installations).

LANG=en_US.UTF-8

Result under 2.1.6:

./kinosearch search=München | grep -Po 'M..?nchen'

München

Both Installations are in a UTF-8 environment, Both WIKI are upgraded to use UTF-8 character set. The word "München" is indexed correctly and delivers the appropiate hits. The conversion back to iso8859-1 shows the wanted result, so there must be a conversion to UTF-8 of documents already converted to utf-8 before:

./kinosearch search=München | grep -Po 'M..?nchen' | iconv -f utf-8 -t iso8859-1

München

Any idea, how I can suppress this utf-8 double conversion in Kinosearch?

Thank you very much for your help in advance
Sincerely yours,
Rainer Daeschler

-- RainerDaeschler - 25 Apr 2022

Hi Rainer, Kinosearch isn't supported anymore. SolrPlugin fills that gap. Have a try.

Further note that Foswiki 2.1.7 has been released with some important security fixes.

-- MichaelDaum - 25 Apr 2022

Thanks a lot. I was thinking of SolrPlugin, but wasn't sure it does search within attachments.

-- RainerDaeschler - 25 Apr 2022

Commenting is disabled while not logged in