Arnon Weinberg
2012-02-26 03:36:20 UTC
This problem seems like it should be common, but I haven't found it
posted anywhere.
Developers may accidentally delete a CVS working directory (ie, the one
with the Entries, Repositories, Root files), and not notice that they
have done so. The cvs command just ignores those directories from then
on, so they don't get updated. When checking out / updating a large
tree, developers may not notice that this directory is then skipped.
? /home/user/directory/pwd
...
Running cvs checkout (or cvs update) without errors or change flags
generally is assumed to mean that the tree is now up to date. However,
in this case, the directory is just skipped (cvs treats it like any file
it doesn't recognize, even though it is in the repository), and the tree
is not sync'ed.
1) How can missing CVS directories be detected?
2) How can they be restored / regenerated?
Arnon Weinberg
www.back2front.ca
posted anywhere.
Developers may accidentally delete a CVS working directory (ie, the one
with the Entries, Repositories, Root files), and not notice that they
have done so. The cvs command just ignores those directories from then
on, so they don't get updated. When checking out / updating a large
tree, developers may not notice that this directory is then skipped.
cd ~/directory/pwd
mv CVS CVS.bak
cvs checkout -A -d ~/directory module
...mv CVS CVS.bak
cvs checkout -A -d ~/directory module
? /home/user/directory/pwd
...
Running cvs checkout (or cvs update) without errors or change flags
generally is assumed to mean that the tree is now up to date. However,
in this case, the directory is just skipped (cvs treats it like any file
it doesn't recognize, even though it is in the repository), and the tree
is not sync'ed.
1) How can missing CVS directories be detected?
2) How can they be restored / regenerated?
Arnon Weinberg
www.back2front.ca