Git-based History Hook for Taskwarrior
Keeping taskwarrior's data under version control grants various advantages, including a detailed change history and of course synchronization capabilities. There are several scripts which aim to provide that functionality but most of them have some little drawbacks. For example, taskgit, task-git.sh and gtw are implemented as wrapper scripts which imply either aliasing or using an unfamiliar command. On the other hand, on-exit_git.py already exploits taskwarrior's hook capabilities but does not provide useful commit messages while maintaining redundant data. This approach also uses taskwarrior's hook mechanism but maintains the data in a more clever manner.
Concept
Instead of tracking all data files only the pending.data file is kept under
version control. The idea is to generate the required information for the
backlog.data, completed.data and undo.data files from the commit history.
For example, the completed.data file contains all deleted and completed
entries which are moved from the pending.data file each time the garbage
collection is invoked [1]. This appears as deletion in the commit history
and thus can easily appended to the unsynced completed.data file on a remote
computer. As mentioned, this can be done for the other two files in a similar
way. By means of the post-merge git hook these three files can be updated
automatically.
Advantages
- The repository can be kept at a minimum size since all information are given by the commit history
- Changes in the history of the
pending.datafile can be projected directly onto the other files
Drawbacks
- If the format of any file changes the generation procedure is likely to break
- Updating the data files after a large git update can be time consuming