Currently, errors that occur in derivations and reactors will throw on the Atom#set that "caused" the error. This is because of the synchronous nature of Sherlock.
It should be possible to set an error handler on reactors to prevent this behavior and to allow handling of errors without disrupting the rest of the application.
This is also needed to be able to reject Promises on any error, see issue #59.
Currently, errors that occur in derivations and reactors will throw on the
Atom#set
that "caused" the error. This is because of the synchronous nature of Sherlock.It should be possible to set an error handler on reactors to prevent this behavior and to allow handling of errors without disrupting the rest of the application.
This is also needed to be able to reject Promises on any error, see issue #59.