fstrube/frak

A lightweight deployment tool that uses rsync to push file changes to remote servers.

frak's Issues

Cannot delete single files

When trying to delete a single file via the path= option, instead the entire project is deployed.

Given the folder:

path/to/project
  ↳ some-file.txt
  ↳ some-folder/
    ↳ another-file.txt

Actual

$ touch new-file.txt
$ mkdir new-folder
$ rm some-file.txt
$ frak path=some-file.txt
*deleting new-file.txt
<f+++++++ new-file.txt
cd+++++++ new-folder/
The above actions will be taken. Continue? (This cannot be undone):

Expected

$ touch new-file.txt
$ mkdir new-folder
$ rm some-file.txt
$ frak path=some-file.txt
*deleting new-file.txt
The above actions will be taken. Continue? (This cannot be undone):

Webhook doesn't work if URL contains a redirect

If the webhook URL redirects then the call fails. This can be very basic, such as HTTP redirecting to HTTPS, or Apache adding a trailing slash to directories.

See this article for a possible solution http://stackoverflow.com/questions/6934185/ruby-net-http-following-redirects.

Diff Not Functioning Properly

Running frak on Ubuntu 14.04.

Returns the error: sed: can't read s/^\(+[^+].*\)/\1/: No such file or directory

The `path=` option stopped working

Cannot use the path= option to target which files to push via Frak. As of version 1.3.1 the path= command-line option isn't working properly.

Expected

$ frak path=app
<fcsT.... app/Models/Core/Permission.php
<fcsT.... app/Models/Core/Role.php
<fcsT.... app/Models/Core/User.php
The above actions will be taken. Continue? (This cannot be undone):

Actual

$ frak path=app
<fcsT.... app/Models/Core/Permission.php
<fcsT.... app/Models/Core/Role.php
<fcsT.... app/Models/Core/User.php
<f+++++++ database/migrations/2016_03_11_194924_mysql_install.php
<fcsT.... public/js/scripts.js
The above actions will be taken. Continue? (This cannot be undone): 

Reported by @billdybas @kfnic and others.

The `init` command should create a `.frak` file with sample configuration

Currently frak init will create a .frak and .frakignore file. The .frak file which it creates is useless because you still have to consult the documentation to figure out what to put in there.

It would be more useful if the file(s) were populated with the samples found in the README.md file.

Auto-update should use git pull

The auto-update feature currently replaces the executable file for frak with the latest version from Github. Unfortunately, if frak has been cloned from a Git repository and symlinked into the user's $PATH, then this procedure will dirty the Git workspace. While this may not always cause problems, there is a chance that it could overwrite in-development work.

Instead, the auto-update function should be smart enough to detect that frak resides in a Git repository and, in turn, use Git to pull from upstream.

Actual:

• Auto-update replaces current file, resulting in unstaged (possibly overwritten) changes to the Git repository

Expected:

• Auto-update detects that frak resides in a Git repository, and uses git pull to update to the latest version

Ability to deploy using Git

Currently, Frak has one deployment strategy: rsync. While this works fine for rapid development, sometimes production work is deployed through Git. Therefore, having a deployment type that would use Git to configure and push to remote repositories may be helpful.

Configuration

label: Example Frak Project
method: git
url: [email protected]:fstrube/frak.git
branch: production

Details
method: git - Tells Frak that we're using the Git method of deployment.
url - The URL of the remote Git repository.
branch - (optional) The branch to use. Default is master.