GiellaLT

GiellaLT provides an infrastructure for rule-based language technology aimed at minority and indigenous languages, and streamlines building anything from keyboards to speech technology.

View GiellaLT on GitHub

Page Content

Setting upa multiuser server

This page is a part of the overall Getting started page. It describes the steps you as a system administrator need to do to prepare the server for the work done by the linguists, by installing the Giellatekno/Divvun infrastructure core, and make it available to the linguists.

NB! It is required that the giella-core and all its subdirectories are readable by all users, and shell scripts must be executable (but this should be handled automatically by Subversion). Nothing more is required, nothing less.

To check out the core module and prepare it, do this:

svn co https://gtsvn.uit.no/langtech/trunk/giella-core
cd giella-core
./autogen.sh
./configure
make
sudo make install

The giella-core and the Giellatekno/Divvun infrastructure is under constant development, and it is important that it is kept up-to-date to maintain a functional working environment for the linguists. To keep it up-to-date, make a shell script or cron job that does the following:

That’s it. The linguists can now check out a single language, and the configure script for that language will find the giella-core and everything else that is needed. You can point the linguists to the following page to get directions on how to check out single languages.

NB!! If you get the following error from svn when checking out:

svn: Can't convert string from 'UTF-8' to native encoding:

it is indicative of your filesystem using a filename encoding not compatible with Unicode. This will stop the svn checkout process, leaving you with an incomplete and most likely non-functional working copy. Typically this can happen if you or your sysadmin has set the following:

LC_ALL=C

To fix this, change or redefine the LC_ALL variable in your ~/.profile file to something like:

LC_ALL=en_UK.UTF-8