Gman builds upon git status to generate a numbered file list.
You can then simply refer to the files by their numbers instead of copy/pasting file names and paths.
Pre-configured commands and actions can be applied to the selected file(s) with various shorthands, such as a for git add.
A few quick examples :
g 2 a # `git add` on the file listed second in your git status. Yep, just three characters.
g 3 4 t # run tests on the third and fourth files
g 2-5 d # display diff for files 2 3 4 and 5Installing Gman is fairly simple, all you have to do is save the file and set up an alias to run it. Using "g" as a shorthand for git-related stuff seems like the obvious thing to do.
Just paste the following line wherever you store your aliases (.bashrc .bash_aliases etc.)
alias g='. ~/gman.sh'Then make sure it is executable with chmod :
chmod +x ~/gman.shAdjust the path according to where you saved the script.
Note: Gman has to be run in source mode (.) to enable some actions like cding to the directory of a file.
Calling the script without arguments makes it run a git status.
The output is displayed as a numbered list and stored.
The list is stored in a text file in a dedicated folder in the /tmp directory. This allows you to use Gman simultaneously in multiple repositories without interference. Also, saving it in the /tmp folder doesn't pollute your local repository, or any folder.
Upon calling the script again, with arguments, that stored list is then used.
The list is only regenerated by calling the script without arguments. If there were changes (new files, commits...) in your repo, call the script without any arguments again to refresh the list and make sure you are not performing actions on the wrong files.
| CLI | Details |
|---|---|
 |
Basic usage : specify a file's number, followed by the desired action. |
 |
Number range : in case you are too lazy. One could also have typed g 3 4 5 6 a for the same outcome. |
 |
You can chain instructions, following the "file numbers - action" pattern. In this example, files 1 2 and 3 are staged, file 6 is checked out, and we cd to the folder containing file 5. |
(Prompt on these screenshots is with MagicMonty's bash-git-prompt)
This is just a short list of some of the available actions. For more details, and to customize them or add your own, take a look at the script : I have tried my best to keep my coding style as clean, functional-y and expressive as possible. Contributions for more custom commands are very welcome.
| Command and shorthand | Description |
|---|---|
| add, a | git add |
| ap | git add -p |
| oops | adds file to the last non-pushed commit |
| r, reset | git reset |
| rm | git rm || rm (does a simple rm if the file isn't tracked) adding f to the command adds the -f parameter (eg: g 3 rmf) |
| checkout, co, u | reverts changes to staged or unstaged file |
| diff, d | shows a diff of the file |
| ds | "short" version : compact, less verbose diff output |
| dc | shows a diff --cached of the file |
| dsc, dcs | short and cached |
| sp | git stash push |
| hist | git log -u |
| Command and shorthand | Description |
|---|---|
| lint | run eslint on the file |
| lintfix | run estlint --fix on the file |
| t | Runs jest on the file, or runs the file itself if it is a test file.It is configured for file.js ↔ file.test.js file naming.You might want to change this. |
| tc | jest --coverage |
| tw | jest --watch |
| tu | jest --updateSnapshot |
| tcwu | yep, you can combine ! |
| twuc | yep, in any order ! |
| tco | runs coverage, opens html report in default browser |
Pro tip for the html report : use a bookmarklet or a browser extention to auto-refresh it every few seconds and use tcw in your terminal. Voilà ! You have a live visualization of your line coverage !
| Command and shorthand | Description |
|---|---|
| cat, c | cat of the selected file |
| bat, b | bat awesome cat alternative. Check it out ! |
| cd | cd to the folder containing the file |
| bak | create a duplicate .bak of the file |
| vim, vi | open the file with vim |
| nano, n | open the file with nano |
| vscode, vs | open the file with vscode |
| atom, at | open the file with atom |
| l | ls of the folder containing the file |
| ll | ls -l |
| la | ls -a |
| lla | lal | ls -la |
| Command and shorthand | Description |
|---|---|
| view, v | The default command when none is specified. Outputs a human readable list of the selected files for a quick review before doing something regrettable afterwards. It also copies the file paths to the clipboard (by default with xclip) for easy pasting. |
| print, p | simple output of the file(s) for piping xargsing and $(more)... ;) |
You can of course add new actions to the script. Look at the existing actions and inspire yourself from them. If you come up with some cool stuff, please contribute :)
You can configure some of the actions of the script to suit your needs. Find the user configuration variables at the beginning of the script and fill in your testing or linting commands, typical file extensions, etc.
You can use Gman with reversed arguments order if you feel more comfortable with it.
For example you could type g a 2 n 3 to stage file number 2 and nano file number 3.
Try it out by supplying the -r parameter in first position (g -r a 2 n 3)
If you want this to change permanently, you can edit the line REVERSED=0 in the user config part of the script.
I have built this for my personal use, and to teach myself bash scripting. I am a front-end web developer and it is not my usual language, but I somehow enjoy it.
Many former colleagues have seen me use this tool and asked me to share it with them, and I gladly did. It's more convenient to share over github so... here we are :)
I have improved it and refactored it a few times and am always looking for advice and constructive criticism, so please feel free to contribute or comment about anything.