Skip to content

HTTPS clone URL

Subversion checkout URL

You can clone with HTTPS or Subversion.

Download ZIP
twmail allows you to mail new tasks to your TaskWarrior inbox
Ruby Shell
branch: master

README.md

twmail

twmail allows you to mail new tasks to your TaskWarrior inbox.

Installation

$ gem install twmail

Usage

  1. Install ruby and this gem
  2. If you don't have a ~/.fetchmailrc yet, copy doc/fetchmailrc.sample to ~/.fetchmailrc
  3. Edit ~/.fetchmailrc and adjust mail account settings (the example was made for Google Mail account). If in doubt, consult the fetchmail documentation, e.g. by executing man fetchmailconf in a terminal.

Motivation

I would like to add new tasks to my TaskWarrior inbox from remote places where I don't have immediate access to my personal TaskWarrior database; e.g. from my iPhone, from work (where I don't have access to my personal TaskWarrior installation) or from another computer.

Using eMail for this looks like a great candidate:

  1. I don't want to (or cannot) install TaskWarrior on all the places and machines where I would like to add tasks from. Sending a note as eMail is pretty much universally available.
  2. Many applications were not made for integration with TaskWarrior. But even the dumbest iPhone app can forward text or a URL via eMail.
  3. eMail is asynchronous by design (fire and forget). Even if disconnected from the net, I can send eMail and the system will deliver it on the very next occassion.

What is missing from a TaskWarrior perspective right now is a way to add these mails to a TaskWarrior installation automatically.

Architecture

The simplest solution I could come up with is this:

  1. A dedicated email account is used to collect the tasks.
  2. A script that imports all eMails as new tasks.

As a prerequisite, TaskWarrior is assumed to be installed and configured. With this architecture in place, the functionality is rather simple to implement:

For each mail{
  Transaction{
    * Fetch mail from mailbox
    * Store mail as new task in Taskwarrior
    * Delete mail from mailbox
  }
}

As the word Transaction implies, the whole operation needs to be atomic per mail. No task must be added if fetching a mail went wrong, and no mail must be deleted if storing the task in TaskWarrior failed.

The solution presented here maintains a one-to-one relation between the INBOX of an mail account and the TaskWarrior database.

TODO Describe how to use fetchmail's daemon mode

TODO Describe how to use fetchmail's IMAP IDLE flag

Components

Mail fetching is done with fetchmail, a proven solution available on all major Unices incl. MacOS. It will be configured to use the twmail script as a mail delivery agent (mda), which means nothing more than that fetchmail fetches the mail from the configured account and hands it over to our script. There is no further storage of the received mails except in TaskWarrior.

Error Handling

If our MDA returns non-zero, fetchmail will not assume the message to be processed and it will try again.

TODO Do we need a dedicated dead-letter queue for all mails fetched, but not successfully processed?

Alternatives

One might think of more elaborate applications that do more clever things, but I wanted to create this solution with as few external dependencies as possible. fetchmail is available on all Unices, and who can afford to live without TaskWarrior anyway? I also played with the thought of a central tasks server that receives mail from services like CloudMailIn and auto-adds them to the server, but the result would not be much different (besides being more complex) to the solution presented here: No task will be fetched into TaskWarrior until the machine with the TaskWarrior database is online.

Another alternative would be to convert the email to JSON and use TaskWarrior's import command. This would allow to create and annotate a new task in one step without the bin/task-uuid workaround.

Advanced Usage

Filtering and Routing

Many more advanced use cases like filtering and routing can be implemented on the mail server side. There are plenty of user interfaces for routing eMails based on their subject, sender, body text, etc. The simplest way to integrate these features with twmail is to use IMAP folders. After all filtering and routing, each eMail must end up in a dedicated IMAP folder (by default, all tasks are fetched from the INBOX folder). twmail can then be configured to do different things depending on which IMAP folder a mail came from.

As an example, here is a simple way to route eMails to different projects in TaskWarrior, based on their subject line:

  1. Set up a dedicated IMAP folder for every project you work on, e.g. "Build Bikeshed", "Reading List", "Get Rich Fast"
  2. Configure the mail server to move every mail from INBOX to the
    1. "Build Bikeshed" folder if the mail subject contains "project:Bikeshed"
    2. "Reading List" folder if the mail subject contains "project:Reading"
    3. "Get Rich Fast" folder if the mail subject contains "project:GetRichFast"
  3. Tell twmail to fetch mails from the "Build Bikeshed", "Reading List", and "Get Rich Fast" IMAP folders (in addition to the INBOX):

TODO Continue description ... may need tracepolls for it.

The approach chosen for twmail also addresses SPAM filtering. Handling that remains the responsibility of the mail server. Anything that makes it to the INBOX is treated as task.

Hooks

twmail comes with an advanced implementation that supports hooks. This makes handling incoming mail very simple for someone familiar with shell scripting, and there is no need to edit the twmail scripts in order to customize its behavior.

When fetchmail is configured to use twmail-hook instead of twmail, the script will call the twmail-hook command (must be in the user's $PATH). Within the hook script, the fields of the parsed email are available as environment variables:

TWMAIL_DATE
TWMAIL_MESSAGE_ID
TWMAIL_FROM
TWMAIL_TO
TWMAIL_SUBJECT
TWMAIL_BODY

Have a look at test/helpers/test_hook for a very simple implementation.

If you prefer a hook with a different name, specify it in the TWMAIL_HOOK environment variable in your .fetchmailrc. For example, if your home directory contains a script called taskwarrior-import.sh, edit the mda line to look like this:

mda TWMAIL_HOOK=~/taskwarrior-import.sh twmail-hook

Housekeeping

By default fetchmail will mark retrieved messages as read, but leave them on the server. For housekeeping purposes, it may be desirable to delete messages from the server once they were successfully imported into TaskWarrior.

There are two ways to achieve this:

  1. Create a filter on the server side that deletes all read mail to a dedicated folder (perhaps "Archive" or "Trash"), or simply deletes it.
  2. Run fetchmail with the --nokeep option, which will delete retrieved messages from the server.

Which option to choose depends on the capabilities of your mail server (Google Mail cannot handle mails based on their read status), and on your level of trust in twmail. I recommend leaving mails on the server until you are confident that everything works as expected.

Testing

twmail comes with a basic set of tests. Execute them by running rake in the cloned source repo.

Contributing

  1. Fork it
  2. Create your feature branch (git checkout -b my-new-feature)
  3. Commit your changes (git commit -am 'Added some feature')
  4. Push to the branch (git push origin my-new-feature)
  5. Create new Pull Request
Something went wrong with that request. Please try again.