This page is part of the web mail archives of SRFI 68 from before July 7th, 2015. The new archives for SRFI 68 contain all messages, not just those from before July 7th, 2015.
The condition hierarchy tree doesn't show &i/o-operation-error and its descendant &i/o-operation-not-available-error. They are specified between &i/o-malformed-filename-error and &i/o-file-protection-error, so you might also want to change the order of them in the specification. The layer-specific condition types are also not included in the list. This may be intentional. file-options-include? is specified to return true or #f, not #t or #f as the other predicates in this SRFI (the other predicates are type predicates). reader? exists but writer? does not exist. reader-close and writer-close are not specified to take any arguments. The documentation of make-i/o-buffer refers to a function names writer-write!, but the specification specifies a function named writer-write!! (with two exclamation marks). open-file-reader+writer is not specified to return results equivalent to open-file-reader and open-file-writer. Is this intentional? Also, there is a typo in the third statement of the definition: "This writer and the writer may ...". More documentation on open-u8vector-writer and writer-u8vector is called for. At the very least, create a little section that states explicitly that the created writer uses a private buffer that is unshared with writer-u8vector's result (if these are indeed the semantics). This is a great I/O framework. I wish I knew more, and could comment on the real, technical issues. That said, I'm a little worried about the octets vs. bits decision -- there is a use for bitwise I/O. - Amir