It’s not that kind of breaking change. It’s a change that won’t affect most people. Only those who chose to use a custom location for their media location and chose to set that to a relative path instead of an absolute one which caused the application to have trouble resolving the paths. The change eliminates a bug by preventing people from doing something that was not intended to be supported. So it’s not a “breaking” change necessarily in the sense that they are changing documented functionality. They are eliminating a way that people can misconfigure the application which may in some cases cause the application to break if someone successfully configured the application in this unintended way.
Immich isn’t a library (the main use case for semver is dependencies that will be pulled into other projects) and as far as I know they don’t state that they use semver.
It only affects undocumented behavior, no documented behavior is being broken.
If you want to consider breakage of undocumented / unintended behavior as a major change, then every bug you fix would require a major version bump, since when you fix something you are essentially breaking compatibility for anyone who might have possibly relied on the existence of that unintended behavior.
You can argue the correctness or not of the guidelines put out at semver.org, but I don’t think there’s any room to argue that announcing a 1.x with a change the developers say is a breaking change, which is what Immich have done, fits within the semver.org guidelines.
Breaking changes should warrant a 2.0 version, not a 1.minor version.
Edit: I am basing my comments on https://semver.org/ guidelines
It’s not that kind of breaking change. It’s a change that won’t affect most people. Only those who chose to use a custom location for their media location and chose to set that to a relative path instead of an absolute one which caused the application to have trouble resolving the paths. The change eliminates a bug by preventing people from doing something that was not intended to be supported. So it’s not a “breaking” change necessarily in the sense that they are changing documented functionality. They are eliminating a way that people can misconfigure the application which may in some cases cause the application to break if someone successfully configured the application in this unintended way.
Immich isn’t a library (the main use case for semver is dependencies that will be pulled into other projects) and as far as I know they don’t state that they use semver.
But it is a service that clients connect to via an API.
The API specification is unaffected by this.
It only affects undocumented behavior, no documented behavior is being broken.
If you want to consider breakage of undocumented / unintended behavior as a major change, then every bug you fix would require a major version bump, since when you fix something you are essentially breaking compatibility for anyone who might have possibly relied on the existence of that unintended behavior.
You can argue the correctness or not of the guidelines put out at semver.org, but I don’t think there’s any room to argue that announcing a 1.x with a change the developers say is a breaking change, which is what Immich have done, fits within the semver.org guidelines.