Skip to content

install.md

Latest commit

 

History

History
194 lines (127 loc) · 5.75 KB

install.md

File metadata and controls

194 lines (127 loc) · 5.75 KB
Raw

Usual way to install packages from source:

  • Locate and download the source code (which is usually compressed)
  • Unpack the source code if it is compressed
  • Compile the code
  • Install the resulting executable
  • Set paths to the installation directory

(If you need this just for installing a text editor, look at micro editor without any hassle. See micro editor install without hassle, see text-editor)

The simplest way/code structure to compile a package (with make) is:

  • Goto the directory containing the package's source code.
  • Type ./configure to configure the package for your system.
  • Type make to compile the package.
  • Optionally(but recommended), type make check to run any tests that come with the package.
  • Type make install to install the programs and any data files and documentation.
  • Optionally, type make clean to remove the program binaries and object files from the source code directory

Example

I have shown how to install nano and R below (version of your choice). Installation of nano is written in more detail.

nano

I will show (in detail) how to install nano-text-editor (latest stable version, to be installed somewhere in your $HOME directory, not system-wise) This is important if you want to install something in HPC without root access

## Goto home directory
cd

## First make a directory specifically to download these kind of programs:
mkdir install_dir

## Download the source-code - copy the link from it's official website (https://nano-editor.org/download.php)
wget https://nano-editor.org/dist/v5/nano-5.8.tar.gz 

## Check installed tar-gzipped file and extract (unzip and untar, see notes for details) it:
tar -xvzf nano-5.8.tar.gz
## (x - extract, v - verbose, z - unzip at the same time, f - file)

## Goto the directory
cd nano-5.8

## mkdir new subdirectory(or any other directory where this program will be installed)
mkdir $HOME/install_dir/install_sub

## Configure utility settings with YOUR SPECIFIC DIRECTORY
./configure --prefix=$HOME/install_dir/install_sub

## build the package:
make

## Optional and recommanded check: 
make check 

## Install the package
make install

## (Optional) Clean the mess
make clean

This package is installed in the $HOME/install_dir/install_sub - but inside the bin folder there. Some programs are installed in the exact folder (i.e., not in another bin subfolder.) So, if you type $HOME/install_dir/install_sub/bin/nano it will open nano. However, this is not always convenient.

Here, we would use a trick to open it (this version) by default tl;dr Add this line to your ~/.bashrc or ~/.zshrc or equivalent:

export PATH=$HOME/install_dir/install_sub/bin:$PATH

Details: We would export(prepend) this path of binary file (the bin location) to the system PATH so that when you just type nano, the nano is opened from this new path. For that, type this command in your terminal (each time you log in).

export PATH=$HOME/install_dir/install_sub/bin:$PATH

To refrain yourself from doing this all the sessions, just add this line to the ~/.bashrc or ~/.zshrc or equivalent.

I have meant both unzip and untar by 'extract'.

  • zip/unzip or gzip/gunzip (or equivalent) commands are to zip/unzip/gzip/un-gzip single files.
  • tar is the process of making a directory into one file. Both are compressed into one command here. If you want to split the commands, it would be something like:
gunzip nano-5.8.tar.gz 		## To un-gzip the file.

tar -xvf nano-5.8.tar 		## To unter the 'tar-file' into a proper folder/directory

For statisticians, you might come across many large *.csv data/other files. It is advisable to gzip them to save space. gzip and gunzip are two simple related commands. R/python(using pandas) can read/write them directly.

R

This part is to install R in iowa state cluster from source.

module load gcc
module load pcre2

# assisting modules
module load cairo
module load libtiff
module load icu4c

# download (get link from official website of R, 
# such as right-click and copy-link the latest version from 'https://cloud.r-project.org/')
wget https://mirror.las.iastate.edu/CRAN/src/base/R-4/R-4.1.2.tar.gz
# unzip and untar
tar -xvzf R-4.1.2.tar.gz
# goto the directory
cd R-4.1.2

# configure the path and other things
./configure --prefix=$HOME/.local/ --with-x=no

# install and check
make
make check

Somehow it is not installing in the correct directory. It is installing in the same directory. If the current directory is path/abc, then you can try path/abc/bin/R would run R with new version you installed.

To make this default R, you can use the same trick used in installation of nano: Add this line to your ~/.bashrc or ~/.zshrc or equivalent:

export PATH=path/abc/bin:$PATH

Using pacman for installation (arch or arch-derivative systems like Manjaro):

Remove old installed packages, leave 1:

sudo paccache -rvk1

Clean AUR built packages:

#Remove orphaned libraries
#pamac remove --orphans

#clean cache
pamac clean --build-files

Clean temporary build files (in your home directory)

From https://forum.manjaro.org/t/cleaning-up-and-freeing-disk-space/6703/31

rm -rf ~/{.bundle,.cargo,.cmake,.dotnet,.electron,.electron-gyp,.gem,.gradle,.lazarus,.node-gyp,.npm,.nuget,.nvm,.racket,.rustup,.stack,.yarn} || true
rm -rf ~/.cache/{electron,electron-builder,go-build,node-gyp,pip,yarn} || true
sudo rm -rf ~/go || true

Snap:

If you use snap, there is a high cance of creating huuuge files. One trck is to remove old versions and retain only last two:

sudo snap set system refresh.retain=2

To clean cache of snap

sudo rm /var/lib/snapd/cache/*