- Personal Access Tokens can now also be granted administator privileges.
- Fixed wrong
usersfolder location.
- Fixed a bug where applications could not access the
/api/v1/jobsendpoint using the new personal access tokens.
- Fix faulty personal access token creation.
- Fix equality handling of
CatalogItem.
- Fixed wrong Docker base image.
- Fixed wrong Docker base image.
- Writer addons did not write all resources to the file(s).
- Do not store all tokens in cookie to solve the "Missing parameters: id_token_hint" error because that makes the cookie very large (> 8 kB). Now only the
id_tokenis stored there.
- Fix log out error "Missing parameters: id_token_hint".
- Simplify containerization.
- Simplify containerization.
- Fixed a bug where a buffer was accessed after free.
- Reverse multithreading changes, there was no bug luckily
- Fixed a bug in the mean-polar-degree algorithm which caused the result of be offset by a random value.
- Fixed a multithreading bug affecting the aggregation calculations. This bug very likely caused incorrect aggregation data probably for a long period of time for datasets with many NaN values.
- Git checkout also checks out submodules.
- It is now possible to configure a default file type.
- Show data writer descriptions in UI.
- Add missing env variable.
- The git tag provider now actually uses the specified tag :-/
- The git tag provider now requires the user to specify the path of the .csproj file.
- The Dockerfile still used .NET 6.
- .NET 7.
- MudBlazor as UI framework.
- Improve cancellation.
- Jobs were not cancelled properly.
- Updated the client to reflect the REST API changes.
- Fixed a bug that prevented normal users to delete their refresh token.
- Added a sync client for Python.
- Fixed a bug with incorrect dependency injection.
- There was an option to limit the memory consumption per request but there was no option to limit the total consumption of the buffers of all requests running in parallel. The old option
DataOptions.ReadChunkSizeis deprecated and has been replaced withDataOptions.TotalBufferMemoryConsumption. Nexus will possibly consume more than this value since it only limits the size of allocated buffers. Additionally, Nexus can't control the memory consumption of loaded extensions.
- Fix missing environment variable in Dockerfile.
- Nexus can now restore from pure git repositories and will compile it during the restore process.
- Added the new claim
CanWriteCatalogGroup.
- Fixed a bug which caused access token lifetime to be very short (2 seconds).
- Enable exporting data without writing a zip file to allow efficient pre-aggregation.
- There was a hard coded listen address in the startup code.
This is the first release of Nexus.