Thanks for your answers! I started with the idea of a YAML file because it was easy to manage and already used by SolR configuration, but it was maybe not a good choice. The technical choice for facet configuration must be safe, and avoid the need for libraries to change the configuration of groups or users on their servers. And I am not fond of the idea of 1 or 2 sysprefs (unimarc/marc21) with a big block of YAML data inside. Not user friendly, and risky (oups, I forgot a space...) So from what I read, I see 2 solutions : - forgetting YAML and putting the configuration in a small new sql table. - putting YAML in a syspref, but not modifying it directly, but through a user friendly form on a specific page. I have never seen that in Koha, so I don't know if it will complies to the coding rules... Mathieu Le 24/09/2013 22:24, Mark Tompsett a écrit :

Greetings,

Chown the files: 664, and make the group www-data (or whatever group the script will end up running as). 666 is just evil. Pun intended. Though, limiting exposure of an entire configuration file by only modifying a single system preference is clearly a more secure idea.

GPML, Mark Tompsett

-----Original Message----- From: Mathieu Saby Sent: Tuesday, September 24, 2013 3:31 PM To: Koha Devel Subject: [Koha-devel] writing a file on server from staff interface

Hi for my patch about facets, I currently have 2 problems : 1- how to put the file in /etc/ of installation directory (in discussion) 2- how to allow the perl script (from staff interface) to write or rewrite the yaml configuration file if a librarian decides to change some settings.

If I do nothing special, YAML::DumpFile fails to (re)write the file, because - I suppose - the apache user does not have rights to access the server. So, for the moment, the only way I found to make my config file editable is "chmod 666". But I suppose it is not very safe...

Do you know if there is a safer way to do that? How is managed for example for SolR configuration files?

Regards Mathieu

-- Mathieu Saby Service d'Informatique Documentaire Service Commun de Documentation Université Rennes 2 Téléphone : 02 99 14 12 65 Courriel : mathieu.saby@univ-rennes2.fr

Thanks Jared I'm looking at didyoumean.pl, and that's what I needed! I will try to write something similar in YAML or JSON (any preferences between the two? I know more YAML...) Mathieu Le 24/09/2013 23:02, Jared Camins-Esakov a écrit :

Mathieu,

I think a YAML file is reasonable, but not in user-editable contexts.

- putting YAML in a syspref, but not modifying it directly, but through a user friendly form on a specific page. I have never seen that in Koha, so I don't know if it will complies to the coding rules..

This is how "Did you mean?" is configured. There wasn't enough configuration required to justify a new table, so I just stuck JSON into the syspref.

Regards, Jared

-- Jared Camins-Esakov Bibliographer, C & P Bibliography Services, LLC (phone) +1 (917) 727-3445 (e-mail) jcamins@cpbibliography.com <mailto:jcamins@cpbibliography.com> (web) http://www.cpbibliography.com/

-- Mathieu Saby Service d'Informatique Documentaire Service Commun de Documentation Université Rennes 2 Téléphone : 02 99 14 12 65 Courriel : mathieu.saby@univ-rennes2.fr

Jared Camins-Esakov schreef op di 24-09-2013 om 17:59 [-0400]:

I prefer JSON because I don't ever edit by hand, and JSON is much faster.

In this case, wouldn't a Perl structure make more sense? -- Robin Sheat Catalyst IT Ltd. ✆ +64 4 803 2204 GPG: 5FA7 4B49 1E4D CAA4 4C38 8505 77F5 B724 F871 3BDF