Notes on the Protocol

A number of enhancements are possible. Also see the file TODO in the CVS source distribution, which has further ideas concerning various aspects of CVS, some of which impact the protocol.

• The Modified request could be speeded up by sending diffs rather than entire files. The client would need some way to keep the version of the file which was originally checked out; probably requiring the use of "cvs edit" in this case is the most sensible course (the "cvs edit" could be handled by a package like VC for emacs). This would also allow local operation of cvs diff without arguments.
• The current procedure for cvs update is highly sub-optimal if there are many modified files. One possible alternative would be to have the client send a first request without the contents of every modified file, then have the server tell it what files it needs. Note the server needs to do the what-needs-to-be-updated check twice (or more, if changes in the repository mean it has to ask the client for more files), because it can't keep locks open while waiting for the network. Perhaps this whole thing is irrelevant if there is a multisite capability (as noted in TODO), and therefore the rcsmerge can be done with a repository which is connected via a fast connection.
• The fact that pserver requires an extra network turnaround in order to perform authentication would be nice to avoid. This relates to the issue of reporting errors; probably the clean solution is to defer the error until the client has issued a request which expects a response. To some extent this might relate to the next item (in terms of how easy it is to skip a whole bunch of requests until we get to one that expects a response). I know that the kerberos code doesn't wait in this fashion, but that probably can cause network deadlocks and perhaps future problems running over a transport which is more transaction oriented than TCP. On the other hand I'm not sure it is wise to make the client conduct a lengthy upload only to find there is an authentication failure.
• The protocol uses an extra network turnaround for protocol negotiation (valid-requests). It might be nice to avoid this by having the client be able to send requests and tell the server to ignore them if they are unrecognized (different requests could produce a fatal error if unrecognized). To do this there should be a standard syntax for requests. For example, perhaps all future requests should be a single line, with mechanisms analogous to Argumentx, or several requests working together, to provide greater amounts of information. Or there might be a standard mechanism for counted data (analogous to that used by Modified) or continuation lines (like a generalized Argumentx). It would be useful to compare what HTTP is planning in this area; last I looked they were contemplating something called Protocol Extension Protocol but I haven't looked at the relevant IETF documents in any detail. Obviously, we want something as simple as possible (but no simpler).
• The scrambling algorithm in the CVS client and server actually support more characters than those documented in section Password scrambling algorithm. Someday we are going to either have to document them all (but this is not as easy as it may look, see below), or (gradually and with adequate process) phase out the support for other characters in the CVS implementation. This business of having the feature partly undocumented isn't a desirable state long-term. The problem with documenting other characters is that unless we know what character set is in use, there is no way to make a password portable from one system to another. For example, a with a circle on top might have different encodings in different character sets. It almost works to say that the client picks an arbitrary, unknown character set (indeed, having the CVS client know what character set the user has in mind is a hard problem otherwise), and scrambles according to a certain octet<->octet mapping. There are two problems with this. One is that the protocol has no way to transmit character 10 decimal (linefeed), and the current server and clients have no way to handle 0 decimal (NUL). This may cause problems with certain multibyte character sets, in which octets 10 and 0 will appear in the middle of other characters. The other problem, which is more minor and possibly not worth worrying about, is that someone can type a password on one system and then go to another system which uses a different encoding for the same characters, and have their password not work. The restriction to the ISO646 invariant subset is the best approach for strings which are not particularly significant to users. Passwords are visible enough that this is somewhat doubtful as applied here. ISO646 does, however, have the virtue (!?) of offending everyone. It is easy to say "But the $ is right on people's keyboards! Surely we can't forbid that". From a human factors point of view, that makes quite a bit of sense. The contrary argument, of course, is that a with a circle on top, or some of the characters poorly handled by Unicode, are on someone's keyboard.

Go to the first, previous, next, last section, table of contents.