Claudia Peter writes: > > we moved our cvs-server from solaris to linux and since then we got > corrupted RCS-files. The symptoms are always the same: in the *,v-file > are after the "desc @@" are a lot of x0 (the number of x0 changes).
That certainly *sounds* like the typical NFS corruption -- if you dump the repository file, is the string of zeros an integral number of disk blocks long and on a block boundary? (Disk blocks are usually 512 or 1024 bytes long.) If so, then there's almost certainly a serious bug in the relevant file system code or the disk subsystem. > The files get corrupted when checking out or > tagging. Checkout doesn't modify the repository file so it's impossible for checkout to *cause* corruption, although it will certainly report it if it exists. > Any hint will be helpfull If you can reproduce the problem, it would be helpful to see a repository file both before and after the corruption. -Larry Jones Is it too much to ask for an occasional token gesture of appreciation?! -- Calvin _______________________________________________ Info-cvs mailing list [email protected] http://lists.gnu.org/mailman/listinfo/info-cvs
