A minimal project manager for the terminal.
This script must be sourced, not run in a subshell. See usage for more information.
At present, prm supports zsh
, as well as bash
.
For more information, see the Wiki page on Zsh support.
Ostensibly, prm also works under Cygwin.
If your $EDITOR
is set to a program external to Cygwin (ex: Sublime Text), you
might want to add export prm_use_cygpath=true
to your .bashrc
/.zshrc
to send
the native Windows path to the editor.
Regrettably, fish
is not supported, because of syntax incompatibilities.
See this issue for some details.
However, Fred Deschenes has made a port for fish
that you could check out.
Additionally, Michael Krieger has integrated prm into a workflow for Alfred (OS X).
This program basically lets you CRUD projects. Upon activation, each projects runs its associated start-script; on deactivation, it runs the project stop-script.
These bash-scripts can be used for things like changing directories, setting environment variables, cleanup, etc.
There is basic prompt integration in the form of [PROJECT] <prompt>
, which can be seen in the animated .gif demo above.
You can have several projects active at once in different shells, as prm associates active instances with the shell PID. Currently active projects can be listed (as described in usage).
Dead project instances (i.e. project instances that are still active on shell exit) will be automatically deactivated the next time you run prm – without running their stop-scripts.
For the motivation behind prm, please see the Wiki page on Problem Statements and Design Goals.
Adding and editing projects will open the associated start- and stop-scripts in your editor (as defined by the $EDITOR
environment variable).
A project start-script might for instance look something like this:
# cd to project directory
cd $HOME/src/Python/hello-world
# activate conda env
source activate hello-world
# show current git status
git status
The same project's stop-script might look like this:
# deactivate conda env
source deactivate hello-world
# clean up
rm *.log *.tmp
When you activate a new project, prm automatically stops any active project in the current shell.
When a project is deactivated, prm changes the working directory back to the path you were originally on before starting your first project.
If you often create projects similar to one you already have, you can load custom scripts from your projects' start.sh
and stop.sh
.
For instance, if you'd like some python-based projects to list the number of outdated packages in their conda envs, you can save
# count outdated packages in conda env
echo "$((($(conda search --outdated --names-only | wc -l)-1))) outdated packages in env"
as e.g. conda-list-outdated.sh
in $PRM_DIR/.common/
(this environment variable is detailed in usage).
You can then load this script in your start- and stop-scripts like so:
prm_load conda-list-outdated
Additionally, if you need the name of the currently active project, this is available via the $PRM_ACTIVE_PROJECT
environment variable.
The prm command line arguments are available in start- and stop-scripts, $3
being the first argument after your project name.
All available environment variables are described on this Wiki page.
To install prm: git clone https://github.com/EivindArvesen/prm.git
, or download and extract a release .zip.
In order to work properly, prm must be sourced, not run in a subshell.
The easiest way to do this is probably to add an alias to prm in your ~/.bashrc
(or wherever you keep your aliases), like so:
alias prm=". path/to/prm.sh"
The technical reason for this is succinctly explained in this Wikipedia article.
You must set your $EDITOR
env-var, e.g.
export EDITOR=nano
in your rc-file (like .bashrc
or .zshrc
).
From the help option screen:
usage: prm <option> [<args>] ...
Options:
active List active project instances.
add <project name> Add project(s).
copy <old> <new> Copy project.
edit <project name> Edit project(s).
list List all projects.
remove <project name> Remove project(s).
rename <old> <new> Rename project.
start <project name> Start project.
stop Stop active project.
-h --help Display this information.
-v --version Display version info.
You can set the prm data directory with the $PRM_DIR
environment variable.
By default all prm data is written to ~/.prm
.
You can install shell completions by running bash completions/deploy_completions.sh
from the project root.
Only bash
is supported for now, but zsh
completions are under development.
Feedback is strongly encouraged. If you run into a bug or would like to see a new feature, please open a new issue. In the case of bugs, please mention what shell they occur under (i.e. bash
or zsh
).
Contributions in the form of code (e.g. implementing new features, bug-fixes) are also appreciated. For information about this, see the Wiki page on Contributing.
Pull requests that do not pass the CI tests will not be merged.
The project tests can be found in the tests
directory.
If you have Bats installed, you can run all tests (for both bash
and zsh
) locally by issuing bash tests/run-tests.sh
in the root project folder of prm.
Unit tests can be found in tests/units.bats
.
End-to-end tests can be found in tests/e2e.bats
.
You can run both by via make test
.
This software is released under the terms of the 3-clause New BSD License. See the license file for details.