cck%cucca@COLUMBIA.EDU (Charlie C. Kim) (07/26/86)
The second distribution of the pre-release of CAP, the Columbia University Appletalk Package for Unix, is now available. The major differences from the previous release are: o atp has been cleaned up considerably and is considered completed. o nbp has been extended to allow name registration with a simple (test) name daemon. nbp is considered as subject to revision. o pap has been extended to include server code. The client code has been cleaned up considerably and tightened (closer to published protocol). pap is considered essentially complete; however, the server code will probably ungo some revision in the near future. o unix glue is a little more considerate of sites other than columbia. Also, packets within a subnet don't go out to a bridge. [See abddp.c, ipmap'ed sites may find this problematic]. o gw (kfps) software modified to allow higher appletalk bridge functionality. (e.g. BrRq's cause lkUps to be broadcast on all nets now). [See cap.notes as mentioned below for more info, many people may find this problematic]. o a tiny bit of documentation for pap and atp has been produced o a number of bug fixes for previous release Like the previous pre-release distribution, a decided effort has been made to construct routines that would be "apple"-like in their calling conventions (and in the protocols followed). ATP and NBP follow very closely the corresponding "Inside Appletalk" documentation. PAP diverges a bit due to two factors. First, certain needs arising when allowing multi-fork operations. Second, the revisions of the protocol definitions in the June 1986 "Inside Appletalk" without the corresponding revision of the calling conventions. Further information on the current pre-release is available on CU20B via anonymous FTP in the directory US:[US.CCK.CAP.D2] (note the new directory) in the files RELEASE and CAP.NOTES. Note that we are unable, due to time and resource constraints, to distribute copies via methods other than anonymous FTP. Apologies to those who have sent mail, but have seemingly been ignored: either the press of time or incomplete path information has prevented response. Charlie C. Kim User Services Group Libraries and Center for Computing Activities Bill Schilit Computer Research Facilities Computer Science Department Columbia University