Join GitHub today
Granularity of XD- and XC-type errors #342
Related to issue #341 a more general question comes up to me: Since we have now multiple p:catch to handle different types of errors, does it still make sense to have such molecular errors as XD0011:
Authors might want to have different error mechanisms for non existing documents, not accessible (protected) document or not well-formed documents. The my understanding such a molecular type of error was ok in XProc 1.0 were the basic audience for error message was human. In XProc 3.0 pipelines will consume errors very much, so I think we need more specific codes to make this possible.
One could argue that the pipeline should inspect the error message (in contract to the error code), but this is the path to non interoperable pipelines because different processors will display different error texts and users might use the pipeline in different language environment and want their error message e.h. in spanish.