A git-based locking mechanism for Unity CCD bucket operations.
npm install unity-ccd-lockerA git-based locking mechanism for Unity CCD bucket operations. Can be used to prevent simultaneous operations.
- Node.js version 8.0.0 or higher _(not needed for executable binaries)_.
- You have Git installed and it is in your PATH.
``sh`
npm i -g unity-ccd-locker
You can use yarn etc. as well.
Executable binaries are available for Windows, Mac, and Linux. You don't anything except git installed on your system to use them.
You can download them from the assets section of a Github release.
`
unity-ccd-locker
You can also call ucl instead of unity-ccd-locker.
unity-ccd-locker requires a git reposiory to store the locks. :
1. Create an empty repository.
2. Run unity-ccd-locker init --repoUrl .
3. (Optional) Set up a config file to prevent providing all arguments in the command line every time.
Notes:
- You only need to perform the setup once per locks repository.
- You can use the same locks repository for everything. You can also have multiple locks repositories for different purposes.
- Do not use the locks repositories for any purpose other than unity-ccd-locker usage.unity-ccd-locker clean-init
- It is highly recommended to use brand new empty repositories for setup. However, you can use existing repositories as well. If you are using an existing repository, run instead of unity-ccd-locker init.
After you have a repo set up, then you can work with locks. This is the recommended workflow:
1. Acquire the lock on a bucket.
2. Perform any CCD operations you wish.
3. Release the lock on the bucket.
If your locks repository got corrupted, or if for some reason you want to start over, you can clean it by running unity-ccd-locker clean --repoUrl . This will delete everything in the repository except the main branch and the initial commit.
Notes:
- After cleaning, a locks repository needs to be setup again before usage.
- If you want to clean and re-setup the repository in one go, you can use unity-ccd-locker clean-init instead of running unity-ccd-locker clean followed by unity-ccd-locker init.
| Name | Aliases | Description |
| --- | --- | --- |
| init | i | Initialize the repository as a locks repository. |clean
| | x | Delete everything in a locks repository. |clean-init
| | xi | Perform a clean, followed by an init. |acquire
| | a | Acquire the lock for a bucket. |release
| | r | Release the lock on a bucket. |show
| | s | Print the lock status of a bucket. |
| Name | Aliases | Type | Required | Description |
| --- | --- | --- | --- | --- |
| --help | -h, -? | flag | No | Print help. |--version
| | -v | flag | No | Print version info. |--config
| | -c | string | No | Path to a JSON config file. |--repoUrl
| | -u | string | Yes | The URL of the locks repository. |--projectId
| | -p | string | For lock commands | CCD project ID. |--envId
| | -e | string | For lock commands | CCD environment ID. |--bucketId
| | -b | string | For lock commands | CCD bucket ID. |--message
| | -m | string | For lock commands | The message to associate with the operation. |--force
| | -f | flag | No | Bypass safety checks. |
Lock commands = acquire, release, show
You can optionally have a JSON config file that you can pass with the --config argument that defines some of the arguments in it, so you don't have to write them all the time.
Keys should correspond to argument names without the dashes. This is how it should look like:
`json`
{
"repoUrl": "
"projectId": "
"envId": "
}
Notes:
- Arguments provided in the CLI have a higher priority than the arguments provided in the config file. This means that if the same argument is provided in both, then the CLI argument will be used.
CCD misbehaves when doing simultaneous operations on the same bucket. If you have ever seen one of these errors, then you very well know what I am talking about:
`
Error! The state of the bucket changed unexpectedly. Please retry this operation.
`
You have an entry specified, but the matching file has not been uploaded, or the file size, or md5 hash of what was uploaded does not match
Sometimes simply retrying works. Other times however, the bucket state gets corrupted and no further operations work properly.
This tool provides a way for you to have _locks_ on buckets, just like how some source control software allows you to lock files to prevent other people from changing them, or how an OS provides file locks to prevent files being altered by multiple processes.
MIT License. Refer to the LICENSE.md file.
Copyright (c) 2023 S. Tarık Çetin