Marshall, Joshua writes: > > The investigation I have had shows that the client / server protocol has > the option for Checksums to be set (I read the protocol for 1.9 [1] > because I couldn't find 1.11) but it says that it is optional. Is there a > way to tell whether this is actually being used?
The standard CVS server implementation only sends checksums when updating a file via a patch (to assure that the file afer applying the patch matches the repository file). (The client/server protocol is documented [sic] in doc/cvsclient.texi in the source distribution. You can browse it on-line at www.cvshome.org.) > I looked through the file in the repository (the ,v file) and can't see > any references to checksums etc being in there. I found a webpage [2] > which stated that there aren't any checksums done on the files. Although > what it says is true about backups being long gone if file corruption is > detected, it would be very useful for the developers to know if a file is > corrupted, and at least know that what they have isn't exactly what was > committed. RCS files do not have checksums. What exactly is it that you're worried about: Are you worried about bugs in CVS? Bugs in your file system? Bugs in your hard drive? Malicious users? > Is there a way to get an external program to checksum these files and > store a checksum somewhere so that they can be checked when they are > pulled out of the repository for validity - and is there a way to do this > automatically? Not without an inordinate amount of work -- what is the higher-level goal that you're trying to accomplish? -Larry Jones These things just seem to happen. -- Calvin _______________________________________________ Bug-cvs mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/bug-cvs
