Install PC from scratch

To be able to remotely connect

sudo apt-get install openssh-server

ccache is used to "cache" the result of g++ compilation. The compilation time is greatly reduced.

export CCACHE_VERSION=3.1.9
cd /tmp
wget http://samba.org/ftp/ccache/ccache-${CCACHE_VERSION}.tar.gz
tar xzvf /tmp/ccache-${CCACHE_VERSION}.tar.gz
cd ccache-${CCACHE_VERSION}
./configure && make && sudo make install

You can already use ccache by setting

export CXX='ccache g++'
export CC='ccache gcc'

cmake is quite deprecated in apt-get sources

export CMAKE_VERSION=2.8.10.2
cd /tmp
wget http://www.cmake.org/files/v2.8/cmake-${CMAKE_VERSION}.tar.gz
tar xzvf /tmp/cmake-${CMAKE_VERSION}.tar.gz
cd cmake-${CMAKE_VERSION}
./configure && make && sudo make install

GNU Emacs is an extensible, customizable text editor—and more. At its core is an interpreter for Emacs Lisp, a dialect of the Lisp programming language with extensions to support text editing. Again, since apt-get packages are quite outdated, we download and install a pretty recent version of emacs.

First install X11 development librairies.

sudo apt-get install libgtk2.0-dev libgif-dev libpng12-dev libxpm-dev

Then go ahead by installing emacs.

export EMACS_VERSION=24.2
cd /tmp
wget http://ftpmirror.gnu.org/emacs/emacs-${EMACS_VERSION}.tar.gz
tar xzvf emacs-${EMACS_VERSION}.tar.gz
cd emacs-${EMACS_VERSION}
./configure --with-jpeg=no --with-tiff=no && make && sudo make install

Antigen is a small set of functions that help you easily manage your shell (zsh) plugins, called bundles. The concept is pretty much the same as bundles in a typical vim+pathogen setup.

git clone https://github.com/zsh-users/antigen.git ~/.config/zsh/antigen

wget --no-check-certificate https://raw.github.com/xgarrido/dotfiles/master/zshrc -O ~/.zshrc

After that open a new terminal in order to get all the changes. Antigen will proceed by downloading different zsh bundle sources, especially plugins related to SuperNEMO software.

To get a effective startup configuration for emacs, one can use the starter-kit setup initially started by technomancy and ported to org-mode by Eric Schulte. We will use a slightly different version of the later that can be found https://github.com/xgarrido/emacs-starter-kit.

Since version 24.x, emacs is bundled with a package manager that allows to download emacs-lisp libraries written from several places such as github, for example.

1. Run (do not forget the \)

   $ \emacs
   

   and wait until packages get installed (it can take 1 or 2 minutes depending on your internet connection).

If you are using all the zsh plugins downloaded by antigen, then you got a problem when you launch emacs which tells you that no emacs server are currently running. This is not a bug, this a feature. You can either run emacs by doing

$ \emacs

or

$ relaunch_emacs

In the latest case, a emacs daemon will be launched first and then all the new emacs instances will be opened from the daemon. The interest of such behavior is that emacs will open almost instantanously since all the loading stuff has been done only one time.

You will need some additionnal packages to make everything working well. Below you will find a list of packages to install.

sudo apt-get install libbz2-dev python-dev libglu1-mesa-dev libmotif-dev libxmu-dev zlib1g-dev

All SuperNEMO software is stored and versionned under subversion centralized version control system. Besides its reliability, SVN offers less options and too "centralization" with respect to git. Using git-svn it is possible to use git and to keep the svn repository safe allowing bidirectional operations between the subversion repository and git. The two next items describe the needed steps to achieve such configuration. Of course, if you do not care about git/svn paradigm, you can skip this part and got to next section.

If you have passed the previous steps, you can now use zsh-aggregator utility and the aggregator command. The last thing you have to do is to define the SNAILWARE_PRO_DIR variable and set it to a directory of your choice. For example, to install all aggregators just do

export SNAILWARE_PRO_DIR=~/SuperNEMO/snware
aggregator build all

On some machines, you can choose a different directory where to build aggregator and the place to install them. SNAILWARE_PRO_DIR is still the install directory but you can export the SNAILWARE_BUILD_DIR variable where the aggregator will be compiled and built. For example, on nemo machines, you can set-up aggregators like that

export SNAILWARE_BUILD_DIR=/scratch/${USER}/SuperNEMO/snware
export SNAILWARE_PRO_DIR=/exp/nemo/${USER}/SuperNEMO/snware
aggregator build all

Aggregators are fine but if you plan do develop some core or analysis programs you may need more flexibility to handle software components. zsh-snailware provides some commands to help in defining a development environment. Basically, you only need to set the SNAILWARE_DEV_DIR directory and then use the snailware command.

For example, to install Bayeux components i.e. datatools, geomtools … in a unique and well-structured directory, you can do

export SNAILWARE_DEV_DIR=~/SuperNEMO/development
snailware git-checkout bayeux
snailware rebuild bayeux

You can also only install datatools and materials, for example

snailware build datatools materials

In such a way, snailware command is a replacement of aggregator packages. Given that you have have installed one of the component, you can then source it

snailware setup all

as well as you can dump the components status

snailware status all

which gives a nice formated table output of the current status of each components.

1. If you decide to use git-svn, you are supposed to know and "master" git command. If not a nice tutorial about git commands can be found at this address http://try.github.com/.
2. To locally commit you just have to git commit your change by stagging some of the modified files. If you want to commit back to the svn repository you should do git svn rebase to make sure nothing has been updated meanwhile and then do git svn dcommit.
3. The first time you dcommit, you may be in trouble due to username mismatch between the svn username you use when checking out and the username you have when commiting. The error looks like

Access to '/svn/!svn/act/...' forbidden

To overcome this problem you should specify explicitly who is committing by doing

git svn dcommit --username=your-username

2013-06-25 : Actually this methods does not seem to work. One way to get rid of this problem is to checkout/commit/update a svn repository using explicitly your username. Something like

svn co http://someplace.org --username your-username

should work. Then go back to your work and try again to commit through git svn.