Sometimes you may want a location to share a temporary file with the following conditions:

  1. The file should be associated with one of your known public domains. In my case, the domain https://paulojeronimo.com.

  2. The file name could be automatically generated for you, via a script, with a small length and random characters.

  3. The file could be encrypted. Ideally, you would use software like GPG for this.

  4. Only you should be capable of removing the file.

  5. Whenever you add, update or remove a file in this location, a file changes.txt stating the changes were made should be updated.

    1. This file should be audited by the command git log so you can see, at any time, what modifications were made in the place.

This project helps you to achieve the conditions above using GitHub Pages!

The tmp git repository always contains a single commit and many temporary files. This repo can also have Bash scripts and other files that are persistent in another git repository (tmp-history). So, for the tmp repository, don’t expect you can make a clone of it, thinking you’ll pull it later because that won’t work. In technical words:

$ git clone https://github.com/paulojeronimo/tmp # <- it will always works!
$ git pull https://github.com/paulojeronimo/tmp  # <- it may not work!

The files generated in this repository are accessible through the URL https://your-github-name.github.io/tmp/.

In my case, I configured GitHub Pages as my custom domain. So, my tmp repo is available here: https://paulojeronimo.com/tmp.

The SHA256 of the temporary files stored here keeps alive (and constantly updated) in tmp-history/blob/main/changes.txt (on GitHub). So, you can audit if there was any change in the files pointed out by it.

All the files that are not temporary in the tmp repository are also available in the repository tmp-history with a history of changes associated with them.

This is the list of the scripts available (in alphabetical order):

Use cases of the scripts above:

You want to fork this project to start using it

  • Step 1 → Create the Git repositories tmp-history and tmp in your GitHub account. Leave them blank because they will be filled in the next steps.

  • Step 2 → Use the fork.sh script. It will clone the tmp-history from an existing origin. Type:

    $ bash <(curl -s https://raw.githubusercontent.com/paulojeronimo/tmp-history/main/fork.sh)

  • Step 3 → Go to the generated tmp-history directory, created in the step above, and call the script generate-tmp.sh

    $ cd tmp-history
$ ./generate-tmp.sh

  • Step 4 → Start using your tmp directory to store temporary or persistent files. Remember: persistent files need to be manually copied to the tmp-history repository. So their history will be kept.

You want to generate a random file (in tmp) to use it in your script 
$ ./generate-random-file.sh

To specify the size (length) of the file name use:

$ size=5 ./generate-random-file.sh

To remove the generated temporary files that have no content, type:

$ find . -maxdepth 1 -type f -size 0 -delete
You want to publish the contents of your directory tmp 
$ ./publish-to-gh-pages.sh
You can turn some files persistent by copying them from tmp to tmp-history

This way they will no longer be temporary and your change history will remain there (in tmp-history).

$ cp file1 file2 ../tmp-history
In the update-history.sh script, the history_dir variable also points to a tmp-history inside the parent directory of tmp repository. So, if your tmp-history directory have a different name, you need to configure this variable.
You want to rebuild this page (index.html) using your own variables

Create a file named my-variables.adoc and configure the AsciiDoc attributes uri-my-*. This is a sample of how to do this:

$ cat > my-variables.adoc <<EOF
:uri-my-domain: https://my-user.github.io/
:uri-my-github: https://github.com/my-user
EOF

Rebuild the index.html with this command:

$ ./README.build.sh