Exception chaining: Difference between revisions

Revision as of 13:18, 1 August 2013

Revision as of 12:38, 27 August 2015

Line 1:

'''Exception chaining''', or '''exception wrapping''', is an [[object-oriented programming]] technique of [[exception handling|handling exceptions]] by re-throwing a caught exception after wrapping it inside a new exception. The original exception is saved as a property (such as ''cause'') of the new exception. The idea is that a method should throw exceptions defined at the same abstraction level as the method itself, but without discarding information from the lower levels.

For example, a method to play a movie file might handle exceptions in reading the file by re-throwing them inside an exception of movie playing. The user interface doesn't need to know whether the error occurred during reading chunk of bytes, tryingor calling eof(). or whatever, and was it unexpected end of file, read error or whatever reading error: itIt needs only the exception message extracted from ''cause''. The user interface layer will have its own set of exceptions. The one interested in ''cause'' can see its stack trace during debugging or in proper log.

Throwing the right kind of exceptions is particularly enforced by [[exception handling#Checked exceptions|checked exceptions]] in the [[Java (programming language)|Java programming language]], and starting with language version 1.4 almost all exceptions support chaining.