Building the perfect CLI-only laptop with the Pinebook Pro
Feb. 17, 2021
Networking
These are popular and baked right into the OS.
- nmcli
- nm-tui
speed-test
Went with this snap instead of fast because fast is not available for arm64 and I want to use on my Pinebook Pro. Instead I’m using speed-test.
CLI Management
Fish Shell
Finally, a command line shell for the 90s
This is much better than the standard bash shell. The directions on their site weren’t the way I needed to set it up, however.
First download and install:
$ sudo apt install fish
Can launch with just fish and starts the shell. But if gonna have it, gotta set as default. To do so, first have to add it to the /etc/shells file.
$ echo /usr/bin/fish | sudo tee -a /etc/shells
Then change which shell is used for the system.
$ chsh -s /usr/bin/fish
Terminator
Setting this up as the terminal to open at boot on the Pinebook Pro. Since this laptop still has a desktop environment, I still want it to boot straight to a terminal emulator. I have it set to autolaunch with these commands:
-mto maximize window-x tmuxto execute tmux
Tmux
tmux / tmux
https://tmuxcheatsheet.com/
Reading Popey’s blog, he has a short bash script set to run to spawn windows running specific apps. This is a simple script that I added one option so that each of the windows are named for easy recognition.
#! /bin/bash
tmux new-window mc -n mc
tmux new-window newsboat -n newsboat
The flag -n is to give the window a name. I put this in ~/bin in a file tilted “apps”, then chmod +x to make it executable. Once the terminal opens, just run the script and all the services will spawn.
bpytop
A very nice looking system monitoring tool and available as a snap.
$ sudo snap install bpytop
Files
duf
A disk usage analyzer for CLI that has a clean interface. It is essentially a TUI interface for df and du in one.
Written in Go and has installers for basically every architecture under the sun, including arm64. See all the notes and releases here - muesli/duf - Github.
Run with:
$ duf
Or pass the path/location at launch.
mc
Very simple to use file manager which behaves more like a GUI file manager instead of cd or ls all the time. Works out perfect for navigating Notes and other files from WebDav.
I wish I could use this while in cadaver, but that tool requires to pull documents to edit, then put to replace.
NextCloud WebDav
This is technically possible, but it is so slow it is not worth using. Besides, I don’t want to call the files as I need them, I want specific files to be local. I want the Nextcloud client to sync the files and then I use a file manager or editor to access them.
This is how to setup the webdav connection with Nextcloud.
Accessing Nextcloud files using WebDAV
There is a specialized CLI utility for WebDav called cadaver. In brief testing this works MUCH better. Speed is not an issue using this client. It is in the official repos, just apt install cadaver. To browse a WebDav directory, start like this:
$ cadaver <path to mountpoint>
It will prompt for password.
To skip entering a password every time, create a file with the base URL of the webdav share plus username password. File should be named .netrc and located in the home directory.
$ nano ~/.netrc
machine my.server
login <username>
password <password>
Then change permissions:
$ chmod 600 ~/.netrc
Samba
This is how I access files on my desktop around my LAN. Mostly for accessing files on my ZFS array that isn’t mounted in Nextcloud. To he honest, instead of using cadaver I’ve just been accessing my Nextcloud files on my desktop via Samba. Then the files aren’t local on my Pinebook Pro and I don’t have to worry about the sync.
First, install:
$ sudo apt install cifs-utils
Next, add a folder where the share will be mounted.
$ sudo mkdir -p /mnt/my.desktop/cloud_bay
Now I can mount it with this command:
sudo mount -t cifs -o username=<username> //<local.dns>/cloud_bay /mnt/my.desktop/cloud_bay/
This will mount this one directory that is part of my Samba setup. I don’t want to mount at boot on the Pinebook Pro, which can significantly slow things down. Instead, I just want to make it easy to mount when I need it. So, let’s create an alias to make the command easy, but not save the password and re-enter every time I want to use it.
$ alias smb_cloud_bay='sudo mount -t cifs -o username=<username> //my.server/cloud_bay /mnt/my.desktop/cloud_bay/'
$ funcsave smb_cloud_bay
Will need to make this for each one of the mounts, but will do it as I run into a situation I need access.
vdirsyncer
A good reference guide for setting up and using a CLI calendar: Terminal calendar with Khal
This utility will sync CardDav and CalDAV to local folders, then can use these files with other programs. I am particularly interested in using this to sync CalDAV for CLI calendar and todo applications on my Pinebook Pro. On other desktops I can use GUI apps, plus the Todo.org app on mobile. This means I would have a self-hosted todo solution that is accessible on all platforms.
There is a steep curve to getting everthing setup.
First, install:
$ sudo apt install vdirsyncer
Next, need to setup some folders. First need the location of the configuration file plus status folder:
$ mkdir -p ~/.vdirsyncer/status
Then make a folder where you want the CalDAV info saved. For me:
$ mkdir ~/.config/todoist_corriveau_cloud
Now, need to setup the config file for the program.
$ nano ~/.vdirsyncer/config
Below is my configuration file.
[general]
status_path = "~/.vdirsyncer/status/"
[pair calendar]
a = "todoist_pbp"
b = "todoist_corriveau_cloud"
collections = ["from a", "from b"]
metadata = ["displayname"]
[storage todoist_pbp]
type = "filesystem"
path = "~/.config/todoist_corriveau_cloud"
fileext = ".ics"
[storage todoist_corriveau_cloud]
type = "caldav"
#Can be obtained from nextcloud
url = "https://<local.dns>/remote.php/dav/calendars/<username>/personal/"
username = "<username>"
password = "<password>"
#SSL certificate fingerprint
verify_fingerprint = "<fingerprint ID see command below"
#Verify ssl certificate. Set to false if it is self signed and not installed on local machine
verify = false
To get your server SSL fingerprint, run this command:
$ echo -n | openssl s_client -connect your.calserver.lcl:443 | openssl x509 -noout -fingerprint
This will display the SSL information and in there will the be fingerprint ID. It will be a series of numbers and letters, similar to a IPv6 address.
Once everything is configured, run:
$ vdirsyncer discover
$ vdirsyncer sync
After syncing, you’ll see a series of .ics files in the local storage (listed above as ~/.config/todosit_corriveau_cloud. These can be used with a CalDAV compliant application.
Last, setup cron to run sync so it updates as changes happen across the devices.
$ crontab -e
and add:
*/15 * * * * vdirsyncer sync
Email / Comms
OfflineIMAP
A tool to download all email from IMAP to a folder on the local machine. The configuration options are deep and complex. I have a working version of the file on the Pinebook Pro.
There is one gotcha to pay attention to and it has to do with the snap version. The “home” directory it is operates in and looks for files is in /home/<user>/snap/<number> not the user home directory in $PATH. For example, if you tell it the configuration file is in ~/Mail it is going to look in /home/<user>/snap/299/Mail for the file, not /home/<user>/Mail.
Here are minimum config to get it to work:
[general]
metadata =~/.offlineimap
maxsyncaccounts = 1
ignore-readonly = yes
socktimeout = 60
[mbnames]
enabled = no
filename = ~/Mutt/muttrc.mailboxes
header = "mailboxes "
peritem = "+%(accountname)s/%(folername)s"
sep = " "
footer = "\n"
incremental = no
[Account dom]
localrepository = LocalGmail
remoterepository = Gmail
autorefresh = 15
synclabels = yes
[Repository LocalGmail]
type = MailDir
localfolders = ~/Gmail
sep = .
sync_deletes = yes
[Repository Gmail]
type = IMAP
remotehost = imap.gmail.com
ssl = yes
sslcacertfile = OS-DEFAULT
remoteport = 993
remoteuser = <email address>
remotepassfile = ~/.config/<file_name>
Instead of putting a password directly in the config file, I opted to create a separate file with an app password, then point the offlineIMAP config to the file.
$ mkdir /home/<user>/snap/299/.config
$ nano /home/<user>/snap/299/.config/<file_name>
Paste in the app password, no other info or configuration needed. Then change the permissions of the file.
$ chmod 600 /home/<user>/snap/299/.config/<file_name>
Mutt
The standard for all terminal email clients.
Once again, I ran into troubles with where files should live with a snap. In this case the configuration file .muttrc needs to live in /home/<user>/snap/mutt/common and not in the ../snap/mutt/appnumber folder most other snaps use. After much fiddling I finally figured this out when I ran snap info mutt and it was the only note.
The mutt configuration file isn’t too complex. The biggest thing to remember is to make some other folders and a file that are called in the configuration.
$ mkdir -p /home/<user>/snap/mutt/common/.mutt/cache
$ mkdir /home/<user>/snap/mutt/common/.mutt/bodies
$ touch /home/<user>/snap/mutt/common/.mutt/certificates
Then, write the configuration file.
# About Me
set from = "<email address>"
set realname = "Dom Corriveau"
# My credentials
set smtp_url = "smtps://<user>@smtp.gmail.com"
set smtp_pass = "<app password>"
set imap_user = "<email address"
set imap_pass = "<app password>"
# My mailboxes
set folder = "imaps://imap.gmail.com"
set spoolfile = "imaps://imap.gmail.com/INBOX"
set postponed = "+[Gmail]/Drafts"
set record = "+[Gmail]/Sent Mail"
set trash = "+[Gmail]/Trash"
# Where to put the stuff
set header_cache = "~/.mutt/cache/headers"
set message_cachedir = "~/.mutt/cache/bodies"
set certificate_file = "~/.mutt/certificates"
# Etc
set ssl_force_tls = yes
set mail_check = 30
set move = no
set imap_keepalive = 900
set sort = threads
set editor = "nano"
set sort = "reverse-date-received"
Todoman
This is the CLI todo application I’ve started with. I would prefer it to be a curses-style application that I can just leave in an open tab, but this is okay because it works with Nextcloud Tasks, which writes the tasks to a calendar. This means can sync using any CalDAV client. On the CLI I am using vdirsyncer to pull down the ICS files for CalDAV, then using Todoman to manage on Pinebook Pro.
First, get Todoman installed and setup the configuration folder and file:
$ sudo apt install todoman
$ mkdir ~/.config/todoman
$ touch ~/.config/todoman/todoman.conf
Now, edit the configuration file - $ nano ~/.config/todoman/todoman.conf and point it at the directory vdirsyncer is also writing files to.
[main]
# A glob expression which matches all directories relevant.
path = ~/.config/todoist_corriveau_cloud/*
date_format = %Y-%m-%d
time_format = %H:%M
default_list = Personal
default_due = 48
That’s it. When todoman is run, it will use the .ics files that are in the folder and when new tasks are added by todoman it will create an `.ics. file and be accessible in all the other notes apps, including Nextcloud Notes, Tasks.og, and Gnome To Do.
There is an awesome tabbed interface that makes entering and editing tasks much easier. First need to install click-repl:
$ pip3 install --user click-repl
Then, when launching the app, use the repl flag to enter the new interface.
$ todoman repl
This interface will give you prompts for start and due date, priority, description, etc.. Once in the click-repl interface you can even run new --interactive and it will walk you through all the options for adding a task.
calcurse
This is the best looking TUI application for calendar. But, I just couldn’t wrap my head around how to configure it to work with a CalDAV setup. Took me forever to realize calcursre-caldav wasn’t a separate package, but a utility built into calcurse itself.
Here’s how to set it up with a CalDAV sync. First install it:
$ sudo apt install calcurse
Next, need to setup the folder along with the configuration file. Run it one time so it will create the base folder. Now, go to the directory calcurse uses and add a CalDAV directory.
$ mkdir ~/.config/calcurse/caldav
$ nano ~/.config/calcurse/caldav/config
Here is my config file:
# If you want to synchronize calcurse with a CalDAV server using
# calcurse-caldav, create a new directory at $XDG_CONFIG_HOME/calcurse/caldav/
# (~/.config/calcurse/caldav/) and $XDG_DATA_HOME/calcurse/caldav/
# (~/.local/share/calcurse/caldav/) and copy this file to
# $XDG_CONFIG_HOME/calcurse/caldav/config and adjust the configuration below.
# Alternatively, if using ~/.calcurse, create a new directory at
# ~/.calcurse/caldav/ and copy this file to ~/.calcurse/caldav/config and adjust
# the configuration file below.
[General]
# Path to the calcurse binary that is used for importing/exporting items.
Binary = calcurse
# Host name of the server that hosts CalDAV.
Hostname = my.server
# Path to the CalDAV calendar on the host specified above.
Path = /remote.php/dav/calendars/<user>/<calendar_name/
# Type of authentication to use. Must be "basic" or "oauth2"
#AuthMethod = basic
# Enable this if you want to skip SSL certificate checks.
InsecureSSL = Yes
HTTPS = Yes
# This option allows you to filter the types of tasks synced. To this end, the
# value of this option should be a comma-separated list of item types, where
# each item type is either "event", "apt", "recur-event", "recur-apt", "todo",
# "recur" or "cal". Note that the comma-separated list must not contain any
# spaces. Refer to the documentation of the --filter-type command line argument
# of calcurse for more details. Set this option to "cal" if the configured
# CalDAV server doesn't support tasks, such as is the case with Google
# Calendar.
# SyncFilter = cal,todo
# Disable this option to actually enable synchronization. If it is enabled,
# nothing is actually written to the server or to the local data files. If you
# combine DryRun = Yes with Verbose = Yes, you get a log of what would have
# happened with this option disabled.
DryRun = No
# Credentials for HTTP Basic Authentication. Leave this commented out if you do
# not want to use authentication.
[Auth]
Username = <username>
Password = <password>
Then, run this to import the info before opening calcurse:
$ calcurse-caldav --init=keep-remote
Now can open calcurse and all is there. It is possbile to run multiple instances of calcurse for separate calendars, but won’t show both together.
Last I setup a crontab to sync the calendar once an hour using the command above. Just add this line to cron:
0 * * * * calcurse-caldav --init=keep-remote
The official documenation for calcurs-caldav:
Entertainment
straw-viewer
A YouTube CLI client that also has a GTK frontend. I’ve used this before for the frontend to avoid going to the YouTube website. This time I’m only using the CLI client and mostly so I can search for YouTube videos from CLI.
This is written in perl, so had to install a bunch of dependencies.
Install cpan to then install perl dependencies and one other program:
$ sudo apt install cpan
$ sudo apt install libwww-perl
Then start cpan:
$ sudo cpan
Which will drop you to a line and is ready to install perl dependencies.
cpan[1]> install Module::Build
cpan[2]> install LWP::Protocol::https
cpan[3]> install Data::Dump
cpan[4]> install JSON
cpan[5]> install LWP::UserAgent:Cached
Now clone the repo:
$ git clone https://github.com/trizen/straw-viewer
Change into the directory and build:
$ perl Build.PL
$ sudo ./Build installdeps
$ sudo ./Build install
The original version of this client required each instance to have a YouTube API key created and inserted. I switch to this version because it uses the invidio.us API.
I only want to play audio version of videos as I mostly use it to listen to new Retro/Synthwave music. To change it to audio only, start it with:
$ straw-viewer --resolution audio
Rather than typing this each time, I just set up an alias.
$ alias strawyt='straw-viewer --resolution audio
$ funcsave strawyt
Runs great and I’m very happy to have this on my Pinebook Pro. If I do want to watch a video, can still start it with straw-viewer and when selecting a video it will open in MPV.
Castero
A terminal podcatcher and podcast player. The key I was looking for was a way to stream podcasts and not just a downloader. Castero will do both with a pretty good looking interface. It is a Python app, but has support for arm64.
To install:
$ pip3 install --user castero
Before launching, there are a couple things to do. First, adjust the playback settings to use VLC as the default. Do this by editing the config file ~/.config/castero/castero.conf. Next, need to set the pip3 install location into $PATH. I do this globally:
$ sudo nano /etc/environment
And add:
/home/<user>/.local/bin
Last, need to import the opml file.
$ castero --import /path/opml_file
ncspot
Easy enough. Install then run and login. Writes the username/password to file so then any terminal that launches it will automatically login.
Passwords
KeePassXC has a built in CLI client, which can be installed even on a headless system. This does require xclip to be installed as well.
Once installed, invoke with:
$ keepassxc-cli
This has very basic function and does not keep the database unlocked. But, it does work.
Step one is to search for the password. To do this use the locate command along with a keyword. The order goes like this:
keepassxc-cli locate <keepass-database-name> <search term>
This should return a result like:
/General/todoist.com
Once you have found the entry, use the clip command to copy the password to the clipboard. If multiple entries exist, but in different directories, put the directory ahead of the name.
keepassxc-cli clip <keepass-database-name> <entry name>
With the example above, you would clip like this:
keepassxc-cli clip keebass-2021-02-01.kbdx /General/todoist.com
Be sure to get the camel-case and syntax correct. Watch out for capital letters in a name that might return negative results just due to camel-case. Put double quotes around anything with spaces to clip the password.
Official repo vs. Snap
In the beginning I tested out using the version in the official repos for Armbian. This version is several releases old and was missing a few features that make using the CLI version of KeePassXC much easier to use. The biggest difference is the ability to open a database and locate/clip passwords without having to enter the master password each time. The Snap is an officially supported version.
Installing the snap:
snap install keepassxc
Starting the CLI client and opening a database:
keepassxc.cli open <database-name>
Once this opens, it drops to a new line and then you can use all the same commands, but don’t have to authorize each time.
Locating:
keybase-20160730> locate todoist
Clipping
keybase-20160730> clip /General/todoist.com
News
I could not find a ncurses feed reader client that also supports FreshRSS connections. Instead, just using Newsboat and importing the OPML. Newsboat is in the repos and to import the OMPL just do:
$ newsboat -I <file>
You can manually add new RSS feeds by editing the “urls” file.
$ nano ~/snap/<number>/.newsboat/urls
Weather
wttr.in to get nice looking weather on CLI very fast. Easy to invoke, no installation needed, just run:
$ curl wttr.in
To call a specific city, run:
$ curl wttr.in/yelm
Instead of typing this every time, set up an alias in bash to run.
$ nano ~/.bashrc
Then add:
alias wttr='curl wttr.in/yelm'
Since I am running fish shell as my primary on the PBP, setting up an alias is different. Rather than editing the ~/.bashrc file, just do:
$ alias wttr='curl wttr.in/yelm'
$ funcsave wttr
and done.
Remaining apps & Research
- Twitter (might have figured out Rainbow Stream, notes coming soon)
- Discord (unlikely)
- XMPP
Contact me
Would love to hear from you on this topic!
Email: obscurednarration@gmail.com
Twitter: @domcorriveau
Get more
These blogs are just the tip of the iceberg. To go deeper and hear more stories around the tech and life around these journals, check out the DC Tech Talks podcast.
Podcast RSS
Apple Podcasts
Spotify
Need even more?
Check out my sister podcast Go With The Heat. We just relaunched the show after going through all of the 80’s goodness that was Miami Vice. Now you’ll find the best of the best from the greatest era of action movies.
Digital marketing professional
Into digital marketing? Get a newsletter all about the best stuff in digital marketing over the last week plus deep-dives into this industry.