From Git to Fossil
People at Git has started to work on a proposal to make the Rust1 programming language mandatory. I don't like Rust and, above all, I don't like its community of little extremist characters who are trying to make everyone swallow their crap by rewriting projects that have been working for decades, doing social media brigading, and other nice little gems worthy of any tiny group with totalitarian delusions. That's why when I see that a project aims to "force" the use of or the switch from C to Rust, to the extent of my possibilities, I flee from it as if I were pursued by the Balrog of the Lord of the Rings with his whip.
I'm old enough to have used (or tested) many version control systems: RCS (Revision Control System), CVS (Concurrent Version System), SVN (Subversion), HG (Mercurial), BZR (Bazaar), and the aforementioned Git, which means I have no problem in switching again, so I started to think about a Git substitute for my personal projects.
The options were to go back to one of the already known or look at something else, and I remembered Fossil. I started looking over the source code and reading the official documentation2 to learn its features, its dependencies, how to install and configure it, etc., and I noticed it had many things I like:
- It's made in C (although it uses some js and tcl for its web functionality) and is a small and efficient program that consumes few hardware resources.
- It's simpler to use and it feels more natural (at least for someone who knows other systems such as Subversion) because it does not contain overkill functionalities such as the staging area, which may be useful in large and complex projects such as the Linux kernel, but for me they have no practical use.
- It allows you to self-host a server on your own quickly and easily because it is already prepared for it. In fact it has a web server and a web interface with version control views, wiki, tickets, etc. On the other hand, with Git you have to use an external software such as GitLab, Gitea or Forgejo, and each one of them is at least one extra software dependency.
- Commit messages don't use email addresses, they use user names so if you have a public repository you don't have to be worried about spam and you don't need a specific email address for this use only.
- It's interoperable with Git in the sense that if there is a need to change a repository back to Git you can do it and it also supports two-way synchronization between a Fossil repository and a Git3 one (this is the functionality used by Fossil and Sqlite projects to manage their GitHub mirrors).
Taking all this into account, I decided to install Fossil and switch my projects from Git. Below you'll see how to do it.
Installing Fossil
The first step is to install Fossil and it's quite likely that the package manager of your operating system already has it available:
sudo apt install fossil # Devuan GNU+Linux
pkg install fossil # FreeBSD
Once your package manager ends the installation you can check its availability with the following command:
fossil version
The command above should return something like the following:
This is fossil version 2.21 [3c53b6364e] 2023-02-26 19:24:24 UTC
Importing Git repositories
Fossil's documentation4 has the following example to export a Git repository and import it as a Fossil repository:
cd repository
git fast-export --all | fossil import --git repository.fossil
I did it differently because I wanted to adjust some things to have the imported repositories "right". As I have several repositories I created two different folders, one, git-exported, to store the exported repositories and another one, fossils, to store the new Fossil repositories:
mkdir ~/git-exported
mkdir ~/fossils
Now you can get into each Git repository folder and export it:
cd repository
git fast-export --all > ~/git-exported/repository.export
Once you have exported all your Git repositories you should go inside the ~/fossils folder and import them one by one with the command:
fossil import --git \
--rename-master trunk \
--attribute "your@mail.com your_username" \
repository.fossil ~/git-exported/repository.export
The --rename-master trunk option renames your Git master branch as
trunk in your new Fossil repository. Fossil, like other version control
systems such as Subversion, uses trunk as the name of the master branch.
If you are among the unfortunate ones who have their master branch named
as "main" this option is not for you and you should check Fossil's
documentation if you want to rename it.
The --attribute "your@mail.com your_username" option changes the
email address "your@mail.com" from Git commits to "your_username" in
the imported Fossil commits. In Fossil the default username is the same
as the one you're currently using in you operating system. Obviously
the given email address should exist in one or more commit messages.
If you want to change more than one email address you can do it
using several --attribute options:
fossil import --git \
--rename-master trunk \
--attribute "your@mail.com your_username" \
--attribute "rms@gnu.org rms" \
--attribute "linus@kernel.org torvalds" \
repository.fossil ~/git-exported/repository.export
The output of the import command looks like the following:
Rebuilding repository meta-data...
100.0% complete...
Vacuuming... ok
project-id: e64b112b40eb3db188060ddb8deeaa96a6ad3b71
server-id: bfbcd4bf8f0f64eaa2d832ddc3b4af00639624a6
admin-user: your_user (password is "XAxPVZcNQ6")
If you look closely, the output gives you the repository administrator's user and password. Keep it because you'll need it to do certain things on your repository.
Setting Up a Fossil Server
Fossil has an embedded web server so you can take advantage of it to create a self-hosted server quickly and easily. The following method is enough for a system with few users, in a private network that cannot be accessed from the outside. There are different ways to set up a Fossil server and some are better than others depending on your needs, so I recommend you to check out the official documentation.
First you should create a folder to store the Fossil repositories on the server:
mkdir /path/to/your/fossils
Next copy your "fossils" to the folder on the server:
scp *.fossil user@your_server:/path/to/your/fossils/
Now you can start the server with the following command:
fossil server --port 8043 \
--cert /path/to/your_cert.pem \
--pkey /path/to/cert/key.pem \
--repolist /path/to/your/fossils/
If you don't have a valid certificate you can use the follwing command:
fossil server --port 8043 \
--cert unsafe-builtin
--repolist /path/to/your/fossils/
Once the server is running, if you put the following url
https://server_address_or_hostname:8043/ in your browser
you'll access a web page with the list of your repositories.
If you click on one of the repositories listed, you'll see the
repository web page. In the navigation menu you should see a
"login" option. Click on it and use the repository administrator's
username and password. Now you can configure the repository to your
liking. Check out Fossil's documentation to learn more.
If you lost your Fossil repository administrator's password, you can recover it by following these steps:
- Go to the repositories folder in your server
- If you don't have Sqlite installed, install it following the usual method of your operating system.
Open repository.fossil with Sqlite:
sqlite3 repository.fossil
Once inside sqlite execute the following query:
select login,pw from user;
The query above will return something like:
your_user|My3w2jRxt1 anonymous|57EBCBD1AAE663B4 nobody| developer| reader|
The password is the value on the second column (in this example the string My3w2jRxt1).
Exit Sqlite with the following command
.quit
Using Fossil
Here is a very brief guide on how to start working with Fossil. This guide assumes that you know how to use Git, at a very basic level, and it's just a starting point, so I recommend you to read Fossil's documentation and the Fossil Book5.
Getting Help
The help command is essential to see which commands are available and the uses and options of each one of these commands. The help command works just like Git:
fossil --help
fossil command_name --help
Creating and cloning repositories
These are similar to Git's but with differences. In Git it's usual that the repository matches the current working folder. In Fossil this is not so, there is a clear separation between the repository, file repository.fossil, and the working (check-out) folder repository_folder.
Creating a repository
You can create a new repository with the command:
fossil init repository.fossil
This command will create only the repository, i.e. the file repository.fossil. To start working with it you have to "open it". Create a folder, move inside it, and open the repository:
mkdir my_working_folder
cd my_working_folder
fossil open /path/to/repository.fossil
You can also tell Fossil the name of the working folder and if it does not exist Fossil will try to create it:
fossil open --workdir my_working_folder /path/to/repository.fossil
Cloning a repository
You can clone a repository using the command fossil clone followed by the
repository url. The url can be a web url, a ssh url, a file path, etc.:
fossil clone https://host/repository
This command will automatically download the file repository.fossil and it will create, at the same level, the working folder repository_folder. If you don't want it to create the working folder you can do it by using the command like:
fossil clone --no-open https://host/repository
Once the repository is cloned, if you want to send your commits to the remote repository (and you have permissions for it), you should configure inside the working folder the remote fossil repository:
fossil remote https://user@host/repository
The command above will asks you for your user's password and it will also asks if you want to save it for future use.
TIP: Because Fossil has this clear separation between the repository file on one hand and the working folders on the other, I keep all the *.fossil files inside a folder named fossils in my /home and I have the working folders where is most appropiate in each case.
Getting Info
For getting the information about the working folder and repository status there are some differences compared to Git, especially in the name and use of the commands.
Viewing the timeline
In Fossil the fossil time and fossil timeline commands are the
equivalent of Git's git log command. Here are some examples from
the one that provides less information to the one that provides more:
fossil time --oneline # similar to git log --oneline
fossil time
fossil time --medium
fossil time --verbose # similar to git log
Viewing changes in your working folder
To see the changes in your working folder compared to the repository there are several commands and each one of them has different options.
To see which files and folders are not under version control:
fossil extras
To see files that are under version control and have been modified:
fossil changes
To see a combination of the two previous commands:
fossil changes --differ
To see the working folder and repository status in a way more closely to the output of Git's git status command:
fossil status --differ
Viewing the Diff(erences)
To see the differences between the contents of our files in the working folder and what is in the repository, Fossil, like Git, has a diff command:
fossil diff
To see the differences between a specific commit and the working folder:
fossil diff --from 2c26dd6b69 # 2c26dd6b69 is the hashtag of the commit
To see the differences between two specific commits:
fossil diff --from 2c26dd6b69 --to cd086a1045
Fossil's diff command doesn't show colorized diffs. Check out the Wiki if you want colorized diffs6
Getting changes
Fossil has an option called autosync that is enabled by default. This option keeps your local repository synchronized with the remote repository. If you have the autosync option enabled, you can get all the changes from the remote repository (if any) with the command:
fossil update
Otherwise it would be more similar to Git's. You have to do first a pull to get the changes and then an update for these to appear in your working folder:
fossil pull
fossil update
Commiting changes
As Fossil does not have the staging area, the commit is much more likely to version control systems like Subversion than Git.
If you want to make a commit with all the pending changes:
fossil commit
The command above will open an editor for you to enter the commit message, but you can also provide the commit message as an option:
fossil commit -m "My commit message"
You can commit specific files with:
fossil commit file1 file2
fossil commit file1 file2 -m "My commit message"
If you have the autosync option enabled, the commit command will also send your changes to the remote server (if any). Otherwise you'll have to send the changes to the remote server with the command:
fossil push
Adding and deleting files
When you need to put new files under version control you can do it with:
fossil add filename
If you want to remove a file from the version control you can do it with:
fossil delete filename
fossil rm filename
By default the rm and delete commands do not physically delete the file from the file system, they simply mark it as no longer under version control.
There's also a command fossil addremove which adds to the repository all the
files in the working folder that are not under version control and removes
from the repository all the files that are under version control but that
no longer exist in the working folder.
Ignoring files
Configuration is one of the points where there are several important
differences between Fossil and Git, so I recommend you to check out
Fossil's documentation. To ignore files and folder you have to create
the file .fossil-settings/ignore-glob inside the working folder:
cd working_folder/
mkdir .fossil-settings
touch .fossil-settings/ignore-glob
Then edit it using glob patterns7:
build/
3rdparty/
*.o
*/a.out
Branches and Tags
Git's paradigm encourages an intensive use of branches and tags but Fossil's paradigm and features are different so the workflows are different too.
As I don't make much use of branches in my personal projects, I think that in order for you to get an idea of the differences so that you can adapt Fossil's features to your workflow, or your workflow to Fossil's features, you better read the following links from the Fossil Wiki:
Local web interface
I don't want to finish this little Fossil guide without making it clear that you can access and use the web interface without having a Fossil server. To do it, run the following command from within the working folder:
fossil ui
Final Thoughts
I like Fossil very much and, in my opinion, it's an almost perfect mix between a centralized version control system like Subversion and a distributed one like Git. I like how easy is to put it in server mode, its web interface and, above all, the simplicity of its command line and of the most common tasks.
I think it's a great program that has a lot to offer, especially to indie programmers and small groups who are interested in self-hosting solutions. If you are looking for a simple and effective alternative to Git do not hesitate to take a look at Fossil because it may be exactly what you are looking for.