The current implementation is extremely memory intensive as input and output arrays, and platform-specific hashing objects are continuously instantiated and destroyed.
The reuse of input can be done in multiplatform.
The reuse of output and hashing objects needs to be done for each platform, and may not be possible fully for for all platforms (e.g. JS doesn't allow to provide an output buffer)
This will require to switch from the general purpose kryptom to a kerberus-specific hashing implementation
The current implementation is extremely memory intensive as input and output arrays, and platform-specific hashing objects are continuously instantiated and destroyed.
The reuse of input can be done in multiplatform.
The reuse of output and hashing objects needs to be done for each platform, and may not be possible fully for for all platforms (e.g. JS doesn't allow to provide an output buffer)
This will require to switch from the general purpose kryptom to a kerberus-specific hashing implementation