Bashmatic® is a BASH framework, meaning its a collection of BASH functions (500+ of them) that, we hope, make BASH programming easier, more enjoyable, and more importantly, usable due to the focus on providing constant feedback to the user about what is happening, as a script that uses Bashmatic is running.
Bashmatic®'s programming style is heavily influenced by Ruby’s DSL languages. If you take a quick look at the is.sh script, it defines a bunch of DSL functions that can be chained with &&
and ||
to create a compact and self-documenting code like this:
[arrow down]
function bashmatic.auto-update() {
local dir
dir=${1:-${BASHMATIC_HOME}}
is.a-directory "${dir}" && {
file.exists-and-newer-than "${dir}/.last-update" 30 && return 0
( cd ${BASHMATIC_HOME} && \
git.is-it-time-to-update && \
git.sync-remote )
}
}
# check if the function is defined and call it
is.a-function.invoke bashmatic.auto-update "$@"
To use it in your own scripts, you’ll want to first study the Examples provided below, and take advantage of ach module available under lib
.
Final note, - once Bashmatic is installed and loaded by your shell init files, you can type is.<tab><tab>
to see what functions are available to you that start with is
. Each module under lib
typically defines public functions starting with the name of the file. Such as, functions in array.sh
typically start with array.<something>.<action>
Bashmatic® offers a huge range of ever-growing helper functions for running commands, auto-retrying, repeatable, runtime-measuring execution framework with the key function run
. There are helpers for every occasion, from drawing boxes, lines, headers, to showing progress bars, getting user input, installing packages, and much more.
Warning
|
A good portion of the helpers within Bashmatic® are written for OS-X, although many useful functions will also work under linux. Our entire test suite runs on Ubuntu. There is an effort underway to convert Homebrew-specifc functions to OS-neutral helpers such as package.install that would work equally well on linux.
|
Start exploring Bashmatic® below with our examples section. When you are ready, the complete entire set of pubic functions (nearly 500 of those) can be found in the functions index page.
And, finally, don’t worry, Bashmatic® is totally open source and free to use and extend. We just like the way it looks with a little ® :)
Warning
|
You can also download the PDF version of this document which is better for print. |
This project was born out of a simple realization made by several very senior and highly experienced engineers, that:
-
It is often easier to use BASH for writing things like universal installers, a.k.a. setup scripts, uploaders, wrappers for all sorts of functionality, such as NPM, rbenv, installing gems, rubies, using AWS, deploying code, etc.
-
BASH function’s return values lend themselves nicely to a compact DSL (domain specific language) where multiple functions can be chained by logical AND
&&
and OR||
to provide a very compact execution logic. Most importantly, we think that this logic is extremely easy to read and understand.
Despite the above points, it is also generally accepted that:
-
A lot of BASH scripts are very poorly written and hard to read and understand.
-
It’s often difficult to understand what the hell is going on while the script is running, because either its not outputting anything useful, OR it’s outputting way too much.
-
When BASH errors occur, shit generally hits the fan and someone decides that they should rewrite the 20-line BASH script in C++ or Go, because, well, it’s a goddamn BASH script and it ain’t working.
Important
|
Bashmatic's goal is to make BASH programming both fun, consistent, and provide plenty of visible output to the user so that there is no mystery as to what is going on. |
Perhaps the easiest way to install Bashmatic® is using this boot-strapping script.
First, make sure that you have Curl installed, run which curl
to see. Then copy/paste this command into your Terminal.
Note
|
The shortcut link resolves to the HEAD version of the bin/bashmatic-install script in Bashmatic Repo.
|
[arrow down]
bash -c "$(curl -fsSL https://bashmatic.re1.re); \
bashmatic-install"
You can pass additional flags to the bashmatic-install
function, including:
-
-v
or--verbose
for displaying additional output, or the opposite: -
-q
or--quiet
for no output -
-l
or--skip-on-login
to NOT install the hook that loads Bashmatic on login. -
If you prefer to install Bashmatic in a non-standard location (the default is
~/.bashmatic
), you can use the-H PATH
flag
For instance, here is a verbose installation with a custom destination:
bash -c "$(curl -fsSL https://bashmatic.re1.re); \
bashmatic-install -v -H ~/workspace/bashmatic"
Here is the complete list of options accepted by the installer:
❯ bin/bashmatic-install --help
USAGE:
bin/bashmatic-install [ flags ]
DESCRIPTION:
Install Bashmatic, and on OSX also installs build tools, brew and latest bash
into /usr/local/bin/bash.
FLAGS:
-H, --bashmatic-home PATH Install bashmatic into PATH (default: ~/.bashmatic)
-V, --bash-version VERSION Install BASH VERSION (default: 5.1-rc2)
-P, --bash-prefix PATH Install BASH into PATH (default: /usr/local)
-l, --skip-on-login Do not install Bashmatic Hook into your dotfiles, which
it does by the default. If you skip it, you can always
change your mind later and add it to your shell dot files
by running the following on the command line:
You can always do so later with the following:
$ ~/.bashmatic/bin/bashmatic load-at-login
This above will install the Bashmatic hook into your shell
dotfile, eg ~/.bash_profile. if you are on BASH,
or ~/.zshrc if you are on ZSH..
-g, --skip-git Do not abort if the destination has local changes
-i, --skip-install Only install/verify prerequisites, skip install.
-v, --verbose See additional output as bootstrap is running.
-q, --quiet See only error output.
-d, --debug Turn on 'set -x' to see all commands running.
-h, --help Show this help message.
When you run bash -c "$(curl -fsSL https://bashmatic.re1.re); bashmatic-install"
, the following typically happens:
-
curl
downloads thebin/bashmatic-install
script and passes it to the built-in BASH for evaluation. -
Once evaluated, function
bashmatic-install
is invoked, which actually performs the installation.-
This is the function that accepts the above listed arguments.
-
-
The script may ask for your password to enable sudo access - this may be required on OS-X to install XCode Developer tools (which include
git
) -
If your version of BASH is 3 or older, the script will download and build from sources version 5+ of BASH, and install it into
/usr/local/bin/bash
. SUDO may be required for this step. -
On OS-X the script will install Homebrew on OS-X, if not already there.
-
Once Brew is installed, brew packages
coreutils
andgnu-sed
are installed, as both are required and are relied upon by Bashmatic.
-
-
The script will then attempt to
git clone
the bashmatic repo into the Bashmatic home folder, or - if it already exists - it willgit pull
latest changes. -
Finally, unless you specify
-l
or--skip-on-login
the script will check your bash dot files, and will add the hook to load Bashmatic from either~/.bashrc
or~/.bash_profile
.
The last part my require some explanation.
Now, you may or may not want to load Bashmatic on login.
In other words, you have something like this in your ~/.bashrc
:
# Let's see if ~/.bashrc mentions Bashmatic:
$ grep bashmatic ~/.bashrc
[[ -f ~/.bashmatic/init.sh ]] && source ~/.bashmatic/init.sh
Then you will have insta-access to all of the convenience functions Bashmtic offers (all 700+ of them).
And while loading the init.sh
file adds time to your login initialzation, and may auto-update the Bashmatic library, there may be good reasons when you prefer NOT to automatically load it on login.
If the above command shows the output you see above, when you grep your bashrc
or zshrc
, then all Bashmatic Functions will be loaded into your shell. This could be very convenient, for instance,
-
you could invoke
ruby.install-ruby-with-readline-and-openssl 3.0.1
to get Ruby installed. -
You could invoke
gem.remote.version sym
to see that the last published verison ofsym
is3.0.1
. -
You could join an array of values with with
array.join ", " apple pear orange
NOTICE: Bashmatic takes no more than 200-300ms to load typically. That said, you might not want to have this many shell functions in your environment, so in that case you can skip login hook by passing -l
or --skip-on-login
.
Install it with:
bash -c "$(curl -fsSL https://bashmatic.re1.re); bashmatic-install -l"
In this case we suggest that you simply add the Bashmatic’s bin
folder to the $PATH
.
For instance:
# ~/.bashrc
export BASHMATIC_HOME="${HOME}/.bashmatic"
export PATH="${BASHMATIC_HOME}/bin:${PATH}"
Then you will have access to the executable script bashmatic
which can be used *as a "gateway" to all bashmatic functions:
You use it like so: bashmatic <function> <args>
:
Important
|
Examples below assume you’ve set the PATH to include ${HOME}/.bashmatic/bin
|
# Eg, if as in the previous example you sourced in Bashmatic:
$ bashmatic.version
2.1.2
# If you have not, you can still invoke 'bashmatic.version':
$ bashmatic version
# Or another function, 'array.join' — if you sourced in init.sh:
$ array.join '|' hello goodbye
hello|goodbye
# Or using the script:
$ bashmatic array.join '|' hello goodbye
hello|goodbye
If you get an error, perhaps Bashmatic® did not properly install.
To discover the breadth of available functions, type the following command to see all imported shell functions:
# List all functions using 4-column mode; print top 5 lines.
❯ bashmatic functions 4 | head -5
7z.a db.psql.connect.db-set hl.yellow-on-gray run.inspect-variables
7z.install db.psql.connect.db-set hr run.inspect-variables-
7z.unzip db.psql.connect.just-d hr.colored run.inspect.set-skip-f
7z.x db.psql.connect.table- http.servers run.on-error.ask-is-en
7z.zip db.psql.connect.table- https.servers run.print-command
# or, to get the count of all functions, use 1 column output:
$ bashmatic functions 1 | wc -l
773
To install Bashmatic manually, follow these steps (feel free to change BASHMATIC_HOME
if you like):
export BASHMATIC_HOME="${HOME}/.bashmatic"
test -d "${BASHMATIC_HOME}" || \
git clone https://github.com/kigster/bashmatic.git "${BASHMATIC_HOME}"
cd "${BASHMATIC_HOME}" && ./bin/bashmatic-install -v
cd ->/dev/null
Sometimes you may not be able to use git
(I have seen issues ranging from local certificate mismatch to old versions of git, and more), but maybe able to download with curl
. In that case, you can lookup the latest tag (substitute "v1.6.0" below with that tag), and then issue this command:
export BASHMATIC_TAG="v1.13.0"
set -e
cd ${HOME}
curl --insecure -fSsl \
https://codeload.github.com/kigster/bashmatic/tar.gz/${BASHMATIC_TAG} \
-o bashmatic.tar.gz
rm -rf .bashmatic && tar xvzf bashmatic.tar.gz && mv bashmatic-${BASHMATIC_TAG} .bashmatic
source ~/.bashmatic/init.sh
cd ${HOME}/.bashmatic && ./bin/bashmatic-install -v
cd ~ >/dev/null
You can always reload Bashmatic® with bashmatic.reload
function. This simply performs the sourcing of ${BASHMATIC_HOME}/init.sh
.
When you install Bashmatic it automatically adds a hook to your ~/.bash_profile
, but if you are on ZSH you may need to add it manually (for now).
Add the following to your ~/.zshrc
file:
[[ -f ~/.bashmatic/init.sh ]] && source ~/.bashmatic/init.sh
Note
|
The entire library takes less than 300ms to load on ZSH and a recent MacBook Pro. |
The top-level Makefile
is mostly provided as a convenience as it encapsulates some common tasks used in development by Bashmatic Author(s), as well as others useful to anyone exploring Bashmatic.
You can run make help
and read the available targets:
❯ make help
docker-build Builds the Docker image with the tooling inside
docker-run-bash Drops you into a BASH session with Bashmatic Loaded
docker-run-fish Drops you into a FISH session with Bashmatic Loaded
docker-run-zsh Drops you into a ZSH session with Bashmatic Loaded
docker-run Drops you into a BASH session
file-stats-git Print all files known to `git ls-files` command
file-stats-local Print all non-test files and run `file` utility on them.
help Prints help message auto-generated from the comments.
install install BashMatic Locally in ~/.bashmatic
open-readme Open README.pdf in the system viewer
release Make a new release named after the latest tag
setup Run the comprehensive development setup on this machine
shell-files Lists every single checked in SHELL file in this repo
tag Tag this commit with .version and push to remote
test-parallel Run the fully auto-g mated test suite
test Run fully automated test suite based on Bats
update-changelog Auto-generate the doc/CHANGELOG (requires GITHUB_TOKEN env var set)
update-functions Auto-generate doc/FUNCTIONS index at doc/FUNCTIONS.adoc/pdf
update-readme Re-generate the PDF version of the README
update-usage Auto-generate doc/USAGE documentation from lib shell files, to doc/USAGE.adoc/pdf
update Runs all of the updates, add locally modiofied files to git.
I’ve added whitespaces around a set of common tasks you might find useful.
Let’s take a quick look at what’s available here.
Makefile is provided as a convenience for running most common tasks and to simplify running some more complex tasks that require remembering many arguments, such as make setup
. You might want to use the Makefile for several reasons:
-
make open-readme
This tasks opens the PDF version of the README in your PDF system viewer.
-
make install
This allows you to install the Bashmatic Framework locally. It simply runs
bin/bashmatic-install
script. At most this will add hooks to your shell init files so that Bashmatic is loaded at login. -
make setup
This task invokes the
bin/dev-setup
script under the hood, so that you can setup your local computer developer setup for software development.Now, this script offers a very rich CLI interface, so you can either run the script directly and have a fine-grained control over what it’s doing, or you can run it with default flags via this make target.
This particular make target runs
bin/dev-setup
script with the following actions:dev, cpp, fonts, gnu, go, java, js, load-balancing, postgres, ruby
-
make test
andmake test-parallel
are both meant for Bashmatic Developers and contributors. Please see the Contributing section on how to run and what to expect from the UNIT tests. -
make update
is the task that should be run by library contributors after they’ve made their their changes and want the auto-generated documentation to reflect the new functions added and so on and so force. This tasks also generates the function index, re-generate the latest PDFs ofREADME
,USAGE
or theCHANGELOG
files.
Note
|
Running make update is is required for submitting any pull request.
|
Bashmatic comes with a Dockerfile that can be used to run tests or jsut manually validate various functionality under linux, and possibly to experiment.
Run make docker-build
to create an docker image bashmatic:latest
.
Run make docker-run-bash
(or …-zsh
or …-fish
) to start a container with your favorite shell, and then validate if your functions work as expected.
Note how this dropped me straight into the Linux environment prompt with Bashmatic already installed.
Why do we need another BASH framework?
BASH is know to be too verbose and unreliable. We beg to differ. This is why we wanted to start this README with a couple of examples.
Just look at this tiny, five-line script:
#!/usr/bin/env bash
source ${BASHMATIC_HOME}/init.sh
h2 "Installing ruby gem sym and brew package curl..." \
"Please standby..."
gem.install "sym" && brew.install.package "curl" && \
success "installed sym ruby gem, version $(gem.version sym)"
Results in this detailed and, let’s be honest, gorgeous ASCII output:
Tell me you are not at all excited to start writing complex installation flows in BASH right away?
Not only you get pretty output, but you can each executed command, it’s exit status, whether it’s been successful (green/red), as well each command’s bloody duration in milliseconds. What’s not to like?!?
Still not convinced?
Take a look at a more comprehensive example next.
In this example, we’ll download and install binaries kubectl
and minikube
binaries into /usr/local/bin
We provided an example script in examples/k8s-installer.sh
. Please click and take a look at the source.
Here is the output of running this script:
Why do we think this type of installer is pretty awesome, compared to a silent but deadly shell script that "Jim-in-the-corner" wrote and now nobody understands?
Because:
-
The script goes out of its way to over-communicate what it does to the user.
-
It allows and reminds about a clean getaway (Ctrl-C)
-
It shares the exact command it runs and its timings so that you can eyeball issues like network congestions or network addresses, etc.
-
It shows in green exit code '0' of each command. Should any of the commands fail, you’ll see it in red.
-
It’s source code is terse, explicit, and easy to read. There is no magic. Just BASH functions.
Note
|
If you need to create a BASH installer, Bashmatic® offers some incredible time savers. |
Let’s get back to the Earth, and talk about how to install Bashmatic, and how to use it in more detail right after.
This final and most feature-rich example is not just an example – it’s a working functioning tool that can be used to install a bunch of developer dependencies on your Apple Laptop.
Note
|
the script relies on Homebrew behind the scenes, and therefore would not work on linux or Windows (unless Brew gets ported there). |
It’s located in bin/dev-setup
and has many CLI flags:
In the example below we’ll use dev-setup
script to install the following:
-
Dev Tools
-
PostgreSQL
-
Redis
-
Memcached
-
Ruby 2.7.1
-
NodeJS/NPM/Yarn
Despite that this is a long list, we can install it all in one command.
We’ll run this from a folder where our application is installed, because then the Ruby Version will be auto-detected from our .ruby-version
file, and in addition to installing all the dependencies the script will also run bundle install
and npm install
(or yarn install
). Not bad, huh?
${BASHMATIC_HOME}/bin/dev-setup \
-g "ruby postgres mysql caching js monitoring" \
-r $(cat .ruby-version) \
-p 9.5 \ # use PostgreSQL version 9.5
-m 5.6 # use MySQL version 5.6
This compact command line installs a ton of things, but don’t take our word for it - run it yourself. Or, at the very least enjoy this one extremely long screenshot :)
This is a great tool that colorizes nearly any other tool''s output.
Run it like so:
${BASHMATIC_HOME}/bin/install-grc
You might need to enter your password for SUDO.
Once it completes, run source ~/.bashrc
(or whatever shell you use), and type something like ls -al
or netstat -rn
or ping 1.1.1.1
and notice how all of the above is nicely colored.
If you are using PostgreSQL, you are in luck! Bashmatic includes numerous helpers for PostreSQL’s CLI
utility psql
.
Note
|
Before you begin, we recommend that you install file .psqlrc from Bashmatic’s conf directory into your home folder. While not required, this file sets up your prompt and various macros for PostgreSQL that will come very handy if you use psql with any regularity.
|
What is db top
anyway?
Just like with the regular top
you can see the "top" resource-consuming processes running on your local system, with dbtop
you can observe a self-refreshing report of the actively running queries on up to three database servers at the same time.
Here is the pixelated screenshot of dbtop
running against two live databases:
In order for this to work, you must first define database connection parameters in a YAML file located at the following PATH: ~/.db/database.yml
.
Here is how the file should be organized (if you ever used Ruby on Rails, the standard config/database.yml
file should be fully compatible):
development:
database: development
username: postgres
host: localhost
password:
staging:
database: staging
username: postgres
host: staging.db.example.com
password:
production:
database: production
username: postgres
host: production.db.example.com
password: "a098098safdaf0998ff79789a798a7sdf"
Given the above file, you should be able to run the following command to see all available (registered in the above YAML file) connections:
$ db connections
development
staging
production
Once that’s working, you should be able run dbtop
:
db top development staging production
Note
|
At the moment, only the default port 5432 is supported. If you are using an alternative port, and as long as it’s shared across the connections you can set the PGPORT environment variable that psql will read.
|
DB Top Configuration:
You can configure the following settings for db top
:
-
You can change the location of the
database.yml
file withdb.config.set-file <filepath>
-
You can change the refresh rate of the
dbtop
with eg.db.top.set-refresh 0.5
(in seconds, fractional values allowed). This sets the sleep time between the screen is fully refreshed.
If you run db
without any arguments, or with -h
you will see the following:
As you might notice, there is an ever-growing list of "actions" — the sub-commands to the db
script.
You can view the full list by passing --commands
flag:
Altgernatively, here is the --examples
view:
You can get a list of all availabled db connections with either
db connections
# OR
db --connections
For instance, a recent addition is the ability to invoke pg_activity Python-based DB "top", a much more advanced top query monitor for PostgreSQL.
You can invoke db pga <connection>
where the connection is taken from the database connection definitions shown above. This is what pg-activity
looks like in action:
Once you know what database you are connecting to, you can then run one of the commands:
- db connect <connection>
-
opens psql session to the given connection
- db db-settings-toml <connection>
-
prints all PostgreSQL settings (obtained with
show all
) as a sorted TOML-formatted file. - db -q list-tables <connection>
-
print a list of all tables in the given database, -q (or --quiet) skips printing the header so that only the table listing is printed.
- db csv <connection> <query>
-
export the result of the query as a CSV to STDOUT, eg
$ db csv filestore "select * from files limit 2"
Results in the following output
component_id,file_path,fingerprint_sha_256,fingerprint_comment_stripped_sha_256,license_info
6121f5b3-d68d-479d-9b83-77e9ca07dd2b,weiboSDK/src/main/java/com/sina/weibo/sdk/openapi/models/Tag.java,
6121f5b3-d68d-479d-9b83-77e9ca07dd2b,weiboSDK/src/main/java/com/sina/weibo/sdk/openapi/models/Comment.java,
Welcome to Bashmatic – an ever growing collection of scripts and mini-bash frameworks for doing all sorts of things quickly and efficiently.
We have adopted the Google Bash Style Guide, and it’s recommended that anyone committing to this repo reads the guides to understand the conventions, gotchas and anti-patterns.
Bashmatic® provides a large number of functions, which are all loaded in your current shell. The functions are split into two fundamental groups:
-
Functions with names beginning with a
.
are considered "private" functions, for example.run.env
and.run.initializer
-
All other functions are considered public.
The following conventions apply to all functions:
-
We use the "dot" for separating namespaces, hence
git.sync
andgem.install
. -
Function names should be self-explanatory and easy to read.
-
DO NOT abbreviate words.
-
All public functions must be written defensively: i.e. if the function is called from the Terminal without any arguments, and it requires arguments, the function must print its usage info and a meaningful error message.
For instance:
$ gem.install
┌─────────────────────────────────────────────────────────┐
│ « ERROR » Error - gem name is required as an argument │
└─────────────────────────────────────────────────────────┘
Now let’s run it properly:
$ gem.install simple-feed
installing simple-feed (latest)...
✔︎ $ gem install simple-feed ▪▪▪▪▪▪▪▪▪▪▪▪▪▪▪▪▪▪▪▪▪〔 5685 ms 〕 0
✔︎ $ gem list > ${BASHMATIC_TEMP}/.gem/gem.list ▪▪▪▪▪▪〔 503 ms 〕 0
The naming convention we use is a derivative of Google’s Bash StyleGuide, using .
to separate BASH function namespaces instead of much more verbose ::
.
After running the above, run bashmatic.functions
function to see all available functions. You can also open the FUNCTIONS.adoc file to see the alphabetized list of all 422 functions.
To get a list of module or pattern-specific functions installed by the framework, run the following:
$ bashmatic.functions-from pattern [ columns ]
For instance:
$ bashmatic.functions-from docker 2
docker.abort-if-down docker.build.container
docker.actions.build docker.containers.clean
.......
docker.actions.update
You can list various modules by listing the lib
sub-directory of the ${BASHMATIC_HOME}
folder.
Note how we use Bashmatic® helper columnize [ columns ]
to display a long list in five columns.
$ ls -1 ${BASHMATIC_HOME}/lib | sed 's/\.sh//g' | columnize 5
7z deploy jemalloc runtime-config time
array dir json runtime trap
audio docker net set url
aws file osx set user
bashmatic ftrace output settings util
brew gem pids shell-set vim
caller git-recurse-updat progress-bar ssh yaml
color git ruby subshell
db sedx run sym
At a high level, the following modules are provided, in order of importance:
The following files provide this functionality:
-
lib/run.sh
-
lib/runtime.sh
-
lib/runtime-config.sh
.
These collectively offer the following functions:
$ bashmatic.functions-from 'run*'
run run.set-next
run.config.detail-is-enabled run.set-next.list
run.config.verbose-is-enabled run.ui.ask
run.inspect run.ui.ask-user-value
run.inspect-variable run.ui.get-user-value
run.inspect-variables run.ui.press-any-key
run.inspect-variables-that-are run.ui.retry-command
run.inspect.set-skip-false-or-blank run.variables-ending-with
run.on-error.ask-is-enabled run.variables-starting-with
run.print-variable run.with.minimum-duration
run.print-variables run.with.ruby-bundle
run.set-all run.with.ruby-bundle-and-output
run.set-all.list
Using these functions you can write powerful shell scripts that display each command they run, it’s status, duration, and can abort on various conditions. You can ask the user to confirm, and you can show a user message and wait for any key pressed to continue.
One of the core tenets of this library is it’s "runtime" framework, which offers a way to run and display commands as they run, while having a fine-grained control over the following:
-
What happens when one of the commands fails? Options include:
-
Ignore and continue (default):
continue-on-error
-
Ask the user if she wants to proceed:
ask-on-error
-
Abort the entire run:
abort-on-error
-
-
How is command output displayed?
-
Is it swallowed for compactness, and only shown if there is an error? (default):
show-output-off
-
Or is it shown regardless?
show-output-on
-
-
Should commands actually run (
dry-run-off
), or simply be printed? (dry-run-on
).
NOTE, in the following examples we assume you installed the library into your project’s folder as
.bashmatic
(a "hidden" folder starting with a dot).
Programming style used in this project lends itself nicely to using a DSL-like approach to shell programming. For example, in order to configure the behavior of the run-time framework (see below) you would run the following command:
#!/usr/bin/env bash
# (See below on the location of .bashmatic and ways to install it)
source ${BASHMATIC_HOME}/init.sh
# configure global behavior of all run() invocations
run.set-all abort-on-error show-output-off
run "git clone https://gthub.com/user/rails-repo rails"
run "cd rails"
run "bundle check || bundle install"
# the following configuration only applies to the next invocation of `run()`
# and then resets back to `off`
run.set-next show-output-on
run "bundle exec rspec"
And most importantly, you can use our fancy UI drawing routines to communicate with the user, which are based on familiar HTML constructs, such as h1
, h2
, hr
, etc.
The lib/output.sh
module does all of the heavy lifting with providing many UI elements, such as frames, boxes, lines, headers, and many more.
Here is the list of functions in this module:
$ bashmatic.functions-from output 3
abort error: left-prefix
ascii-clean h.black ok
box.blue-in-green h.blue okay
box.blue-in-yellow h.green output.color.off
box.green-in-cyan h.red output.color.on
box.green-in-green h.yellow output.is-pipe
box.green-in-magenta h1 output.is-redirect
box.green-in-yellow h1.blue output.is-ssh
box.magenta-in-blue h1.green output.is-terminal
box.magenta-in-green h1.purple output.is-tty
box.red-in-magenta h1.red puts
box.red-in-red h1.yellow reset-color
box.red-in-yellow h2 reset-color:
box.yellow-in-blue h2.green screen-width
box.yellow-in-red h3 screen.height
box.yellow-in-yellow hdr screen.width
br hl.blue shutdown
center hl.desc stderr
columnize hl.green stdout
command-spacer hl.orange success
cursor.at.x hl.subtle test-group
cursor.at.y hl.white-on-orange ui.closer.kind-of-ok
cursor.down hl.white-on-salmon ui.closer.kind-of-ok:
cursor.left hl.yellow ui.closer.not-ok
cursor.rewind hl.yellow-on-gray ui.closer.not-ok:
cursor.right hr ui.closer.ok:
cursor.up hr.colored warn
debug inf warning
duration info warning:
err info:
error left
Note that some function names end with :
– this indicates that the function outputs a new-line in the end. These functions typically exist together with their non-:
-terminated counter-parts. If you use one, eg, inf
, you are then supposed to finish the line by providing an additional output call, most commonly it will be one of ok:
, ui.closer.not-ok:
and ui.closer.kind-of-ok:
.
Here is an example:
function valid-cask() { sleep 1; return 0; }
function verify-cask() {
inf "verifying brew cask ${1}...."
if valid-cask ${1}; then
ok:
else
not-ok:
fi
}
When you run this, you should see something like this:
$ verify-cask TextMate
✔︎ verifying brew cask TextMate....
In the above example, you see the checkbox appear to the left of the text. In fact, it appears a second after, right as sleep 1
returns. This is because this paradigm is meant for wrapping constructs that might succeed or fail.
If we change the valid-cask
function to return a failure:
function valid-cask() { sleep 1; return 1; }
Then this is what we’d see:
$ verify-cask TextMate
✘ verifying brew cask TextMate....
Components are BASH functions that draw something concrete on the screen. For instance, all functions starting with box.
are components, as are h1
, h2
, hr
, br
and more.
$ h1 Hello
┌───────────────────┐
│ Hello │
└───────────────────┘
These are often named after HTML elements, such as hr
, h1
, h2
, etc.
Here is another example where we are deciding whether to print something based on whether the output is a proper terminal (and not a pipe or redirect):
output.is-tty && h1 "Yay For Terminals!" output.has-stdin && echo "We are being piped into..."
The above reads more like a high level language like Ruby or Python than Shell. That’s because BASH is more powerful than most people think.
There is an example script that demonstrates the capabilities of Bashmatic.
If you ran the script, you should see the output shown in this screenshot. Your colors may vary depending on what color scheme and font you use for your terminal.
You can reliably install ruby gems or brew packages with the following syntax:
#!/usr/bin/env bash
source ${BASHMATIC_HOME}/init.sh
h2 "Installing ruby gem sym and brew package curl..." \
gem.install sym
brew.install.package curl
success "installed Sym version $(gem.version sym)"
When you run the above script, you shyould seee the following output:
You can shorten URLs on the command line using Bitly, but for this to work, you must set the following environment variables in your shell init:
export BITLY_LOGIN="<your login>"
export BITLY_API_KEY="<your api key>"
Then you can run it like so:
$ url.shorten https://raw.githubusercontent.com/kigster/bashmatic/master/bin/install
# http://bit.ly/2IIPNE1
There are a couple of Github-specific helpers:
github.clone github.setup
github.org github.validate
For instance:
$ github.clone sym
✘ Validating Github Configuration...
Please enter the name of your Github Organization:
$ kigster
Your github organization was saved in your ~/.gitconfig file.
To change it in the future, run:
$ github.org <org-name>
✔︎ $ git clone [email protected]:kigster/sym ▪▪▪▪▪▪〔 931 ms 〕
$ bashmatic.functions-from file
file.exists_and_newer_than file.list.filter-non-empty
file.gsub file.size
file.install-with-backup file.size.mb
file.last-modified-date file.source-if-exists
file.last-modified-year file.stat
file.list.filter-existing
For instance, file.stat
offers access to the fstat()
C-function:
$ file.stat README.md st_size
22799
$ bashmatic.functions-from array
array.to.bullet-list array.includes
array.has-element array.includes-or-exit
array.to.csv array.from.stdin
array-join array.join
array-piped array.to.piped-list
array.includes-or-complain
For instance:
$ declare -a farm_animals=(chicken duck rooster pig)
$ array.to.bullet-list ${farm_animals[@]}
• chicken
• duck
• rooster
• pig
$ array.includes "duck" "${farm_animals[@]}" && echo Yes || echo No
Yes
$ array.includes "cow" "${farm_animals[@]}" && echo Yes || echo No
No
The utilities module has the following functions:
$ bashmatic.functions-from util
pause.long util.install-direnv
pause util.is-a-function
pause.short util.is-numeric
pause.medium util.is-variable-defined
util.append-to-init-files util.lines-in-folder
util.arch util.remove-from-init-files
util.call-if-function util.shell-init-files
shasum.sha-only util.shell-name
shasum.sha-only-stdin util.ver-to-i
util.functions-starting-with util.whats-installed
util.generate-password watch.ls-al
For example, version helpers can be very handy in automated version detection, sorting and identifying the latest or the oldest versions:
$ util.ver-to-i '12.4.9'
112004009
$ util.i-to-ver $(util.ver-to-i '12.4.9')
12.4.9
Ruby helpers abound:
$ bashmatic.functions-from ruby
bundle.gems-with-c-extensions ruby.install-ruby-with-deps
interrupted ruby.install-upgrade-bundler
ruby.bundler-version ruby.installed-gems
ruby.compiled-with ruby.kigs-gems
ruby.default-gems ruby.linked-libs
ruby.full-version ruby.numeric-version
ruby.gemfile-lock-version ruby.rbenv
ruby.gems ruby.rubygems-update
ruby.gems.install ruby.stop
ruby.gems.uninstall ruby.top-versions
ruby.init ruby.top-versions-as-yaml
ruby.install ruby.validate-version
ruby.install-ruby
From the obvious ruby.install-ruby <version>
to incredibly useful ruby.top-versions <platform>
– which, using rbenv and ruby_build plugin, returns the most recent minor version of each major version upgrade, as well as the YAML version that allows you to pipe the output into your .travis.yml
to test against each major version of Ruby, locked to the very latest update in each.
$ ruby.top-versions
2.0.0-p648
2.1.10
2.2.10
2.3.8
2.4.9
2.5.7
2.6.5
2.7.0
2.8.0-dev
$ ruby.top-versions jruby
jruby-1.5.6
jruby-1.6.8
jruby-1.7.27
jruby-9.0.5.0
jruby-9.1.17.0
jruby-9.2.10.0
$ ruby.top-versions mruby
mruby-dev
mruby-1.0.0
mruby-1.1.0
mruby-1.2.0
mruby-1.3.0
mruby-1.4.1
mruby-2.0.1
mruby-2.1.0
These are fun helpers to assist in scripting gem management.
$ bashmatic.functions-from gem
g-i gem.gemfile.version
g-u gem.global.latest-version
gem.cache-installed gem.global.versions
gem.cache-refresh gem.install
gem.clear-cache gem.is-installed
gem.configure-cache gem.uninstall
gem.ensure-gem-version gem.version
For instance
$ g-i awesome_print
✔︎ gem awesome_print (1.8.0) is already installed
$ gem.version awesome_print
1.8.0
There are plenty more modules, that help with:
-
Ruby Version Helpers and (Ruby Gem Helpers)[lib/gem.sh], that can extract curren gem version from either
Gemfile.lock
or globally installed gem list.. -
AWS helpers – requires
awscli
and credentials setup, and offers some helpers to simplify AWS management. -
Docker Helpers – assist with docker image building and pushing/pulling
And many more.
See the full function index with the function implementation body in the FUNCTIONS.adoc index.
The following example is the actual code from a soon to be integrated AWS credentials install script. This code below checks that a user has a local ~/.aws/credentials
file needed by the awscli
, and in the right INI format. If it doesn’t find it, it checks for the access key CSV file in the ~/Downloads
folder, and converts that if found. Now, if even that is not found, it prompts the user with instructions on how to generate a new key pair on AWS IAM website, and download it locally, thereby quickly converting and installing it as a proper credentials file. Not bad, for a compact BASH script, right? (of course, you are not seeing all of the involved functions, only the public ones).
# define a new function in AWS namespace, related to credentials.
# name of the function is self-explanatory: it validates credentials
# and exits if they are invalid.
aws.credentials.validate-or-exit() {
aws.credentials.are-valid || {
aws.credentials.install-if-missing || bashmatic.exit-or-return 1
}
}
aws.credentials.install-if-missing() {
aws.credentials.are-present || { # if not present
aws.access-key.is-present || aws.access-key.download # attempt to download the key
aws.access-key.is-present && aws.credentials.check-downloads-folder # attempt to find it in ~/Downloads
}
aws.credentials.are-present || { # final check after all attempts to install credentials
error "Unable to find AWS credentials. Please try again." && bashmatic.exit-or-return 1
}
bashmatic.exit-or-return 0
}
Now, how would you use it in a script? Let’s say you need a script to upload something to AWS S3. But before you begin, wouldn’t it be nice to verify that the credentials exist, and if not – help the user install it? Yes it would.
And that is exactly what the code above does, but it looks like a DSL. because it is a DSL.
This script could be your bin/s3-uploader
aws.credentials.validate-or-exit
# if we are here, that means that AWS credentials have been found.
# and we can continue with our script.
Some bash files exists as libraries to be "sourced in", and others exist as scripts to be run. But users won’t always know what is what, and may try to source in a script that should be run, or vice versa - run a script that should be sourced in.
What do you, programmer, do to educate the user about correct usage of your script/library?
Bashmatic® offers a reliable way to test this:
#!/usr/bin/env bash
# load library
if [[ -f "${Bashmatic__Init}" ]]; then source "${Bashmatic__Init}"; else source ${BASHMATIC_HOME}/init.sh; fi
bashmatic.validate-subshell || return 1
If you’rather require a library to be sourced in, but not run, use the code as follows:
#!/usr/bin/env bash
# load library
if [[ -f "${Bashmatic__Init}" ]]; then source "${Bashmatic__Init}"; else source ${BASHMATIC_HOME}/init.sh; fi
bashmatic.validate-sourced-in || exit 1
If you are stuck working on a monitor that does not support switching digit input from TV to PC, NOR does OS-X show the "underscan" slider in the Display Preferences, you may be forced to change the underscan manually. The process is a bit tricky, but we have a helpful script to do that:
$ source init.sh
$ change-underscan 5
This will reduce underscan by 5% compared to the current value. The total value is 10000, and is stored in the file /var/db/.com.apple.iokit.graphics
. The tricky part is determining which of the display entries map to your problem monitor. This is what the script helps with.
Do not forget to restart after the change.
Acknowledgements: the script is an automation of the method offered on this blog post.
Please submit a pull request or at least an issue!
The framework comes with a bunch of automated unit tests based on the fantastic framework bats
.
Bats is auto-installed by the bin/specs
script.
We use Bats framework for testing, however we provided a convenient wrapper bin/specs
which installs Bats and its dependencies so that we don’t have to worry about installing it manually.
The script can be run:
-
Without any arguments to run all tests in the
test
folder, or -
You can pass one or more existing test file paths as arguments, eg
bin/specs test/time_test.bats
-
Finally, you can pass an abbreviated test file name — eg "time" will resolve to
test/time_test.bats
The script accepts a bunch of CLI arguments and flags shown below:
Alternatively, you can run the entire test suite via the Makefile, using one of two targets:
# Sequential
make test
# Parallel
make test-parallel
One of the very useful flags to bin/specs
script is the -p/--parallel
.
If you invoke it with this flag, the script will install GNU Parallel utility, which is in itself worth reading about. We refer you to the following set of YouTube Introductory Videos on taking advantage of GNU Parallel projects and it’s executable.
Below is the screenshot of the tests running with the parallel flag. The script automatically detects that my machine has 16 CPU cores and uses this as a parallization factor.
Note that you can run all tests in less than 15 seconds by using GNU parallel. Just run the following make target, and it will install any dependencies.
make test-parallel
While not every single function is tested (far from it), we do try to add tests to the critical ones.
Please see existing tests for the examples.