Matterhorn is a terminal client for the Mattermost chat system.
New Release Notifications
Get notified about new Matterhorn releases by following our Twitter account!
https://twitter.com/matterhorn_chat
Quick Start
We provide pre-built binary releases for some platforms. Please see the
release list to download a binary release for your platform that matches
your server version:
https://github.com/matterhorn-chat/matterhorn/releases
To fetch a release and run Matterhorn, run the following commands (where
VERSION
and PLATFORM
match your setup):
wget https://github.com/matterhorn-chat/matterhorn/releases/download/<VERSION>/matterhorn-<VERSION>-<PLATFORM>.tar.gz
tar xf matterhorn-<VERSION>-<PLATFORM>.tar.gz
cd matterhorn-<VERSION>-<PLATFORM>
./matterhorn
When you run Matterhorn you'll be prompted for your server information
and credentials. At present matterhorn
supports only username/password
authentication.
Note: Version ABBCC.X.Y
matches Mattermost server version A.BB.CC
.
For example, if your Mattermost server version is 3.6.0
then you
would download matterhorn version 30600.2.4
. See Our Versioning
Scheme for details.
Configuring
For configuration options you have two choices:
- Interactive configuration entered on each program run
- Configuration via stored settings in a config file
The first option is useful when trying out the program because you can
get up and running without worrying about making a configuration. Once
you're ready to make your settings persistent, they can be added to
a configuration file. An example configuration file can be found at
sample-config.ini
. Any settings omitted from the configuration will be
obtained interactively at startup.
When looking for configuration files, matterhorn will prefer
config.ini
in the current working directory, but will look in the
typical XDG configuration directories (you'll probably want to use
$HOME/.config/matterhorn/config.ini
) and as a last resort look for a
globally-accessible /etc/matterhorn/config.ini
.
Using the Client
The user interface has three main areas:
- Left: list of channels you're in, and list of users in your team and
their statuses (
+
means online, -
means away, ×
means Do Not
Disturb, and an absent sigil means offline)
- Right: messages in the current channel
- Bottom: editing area for writing, editing, and replying to messages
You can use built-in keybindings or /cmd
-style commands to operate
the client. To see available keybindings and commands, use the default
binding of F1
or run the /help
command. Keybindings may include
modifiers such as Control (indicated with a C-
prefix) or Meta
(indicated with a M-
prefix). If your keyboard has an Alt
key, that
will work as Meta
. If it does not, you may be able to configure your
terminal to provide Meta
via other means (e.g. iTerm2 on OS X can be
configured to make the left Option key work as Meta).
To join a channel, use the /join
command to choose from a list of
available channels. To create a channel, use /create-channel
. To leave
a channel, use /leave-channel
.
To create a private group chat amongst yourself and other users, use the
/group-msg
command, e.g., /group-msg user1 user2
.
To see the members in the current channel, use the /members
command.
To send a message, type it into the editor and press Enter to send.
To send a multi-line message, toggle multi-line mode with the default
binding M-e
. Markdown syntax is accepted.
To edit your current message in an external editor ($EDITOR
), use the
default binding of M-k
.
To preview the message you're about to send while you compose it (e.g.
to check on how your Markdown syntax will be rendered), toggle preview
mode with the default binding M-p
.
To change channels, use /focus
or one of the default bindings C-n
(next channel), C-p
(previous channel), C-g
(fast channel switch).
To directly message another user, use /focus
or C-g
.
C-g
channel switching mode does a substring match of the input text on
the channel and usernames; metacharacters ^
and $
at the beginning
or end of input, respectively, anchor the match in case of multiple
matches. The cursor in this mode is usable with C-n
and C-p
.
To switch to the channel you were in prior to the current channel, use
the default binding M-s
(swap). The most recent channel is marked in
the channel list with a "<
" indicator.
To switch to the next channel with unread messages, use the default
binding M-a
.
To quickly show a list of URLs mentioned in the current channel and then
open one in your local browser, use the default binding of C-o
and
configure the urlOpenCommand
configuration setting.
To edit, delete, flag, or reply to a message, select a message with
the default binding of C-s
. Use the default binding of C-c
to
cancel these operations.
Messages that have been flagged can be viewed with either the /flags
command or M-8
. This view allows you to select and unflag particular
messages, as well.
To enable spell-checking in the message editor, install Aspell and set
enableAspell
to True
in your configuration. To override Aspell's
choice of master dictionary, set the aspellDictionary
option to the
name of the dictionary you'd like to use.
Features
- Channel creation, deletion, and membership management commands
- Optimized channel-switching modes:
M-a
, M-s
, and C-g
- Message posting, editing, replying, and deletion
- Markdown rendering
- Convenient URL-opening with local browser
- Secure password entry via external command (e.g. OSX keychain)
- Yank verbatim content from messages into the system clipboard
- Preview message rendering before sending
- Optional smart quoting for efficient Markdown entry
- Edit messages with
$EDITOR
- Message editor with kill/yank buffer and readline-style keybindings
- Tab-completion of usernames, channel names, and commands
- Spell-checking via Aspell
- Syntax highlighting of fenced code blocks in messages (works best in
256-color terminals)
- Flagging and unflagging of posts, which are then viewable with
M-8
or /flags
- Support for SOCKS 4 and 5 proxies via the
ALL_PROXY
, HTTP_PROXY
,
and HTTPS_PROXY
environment variables. (Plain HTTP proxies are not
yet supported.)
- Multiple color themes with color theme customization support
Spell Checking Support
Matterhorn uses aspell
to perform spell-checking of your message
input. To use this feature:
Building
matterhorn
is built with the provided install.sh
script, which
requires git
and an appropriate ghc
/cabal
installation.
(Although the name suggests installtion, this will just do a build in
dist-newstyle
.) This script will pull the appropriate repos and build
the application. This is required for building Matterhorn since clones
of some of our other dependencies may need to be locally available in
deps/
in case important changes to those dependencies have not yet
been released.
Our Versioning Scheme
Matterhorn version strings will be of the form ABBCC.X.Y
where ABBCC
corresponds to the Mattermost server version supported by the release.
For example, if a release supports Mattermost server version 1.2.3, the
ABBCC portion of the matterhorn
version will be 10203
. The X.Y
portion of the version corresponds to our own version namespace for the
package. If the server version changes, X.Y
SHOULD be 0.0
. Otherwise
the first component should increment if the package undergoes major code
changes or functionality changes. The second component alone should
change only if the package undergoes security fixes or other bug fixes.
Contributing
If you decide to contribute, that's great! Here are some guidelines you
should consider to make submitting patches easier for all concerned:
- If you want to take on big things, let's have a design/vision
discussion before you start coding. Create a GitHub issue and we can
use that as the place to hash things out. We'll be interested to
discuss any usability / UI, performance, or compatibility issues.
- Please make changes consistent with the conventions already used in
the codebase.
- We follow a few development practices to support our project and it
helps when contributors are aware of these. Please see
PRACTICES.md
for more information.
Frequently Asked Questions
- Q: Does matterhorn support Gitlab authentication?
- A: No. But we would be happy to work with contributors who are
interested in investigating what this would take and/or implementing
it. See the Contributing section for details.