Speaking from blissful ignorance, I would speculate that the entries in .cvsignore must be ordinary files, not directories. CVS is going to navigate the tree anyway, but .cvsignore tells it what to do with the files it finds in each directory.

According to info,

* As CVS traverses through your directories, the contents of any `.cvsignore' will be appended to the list. The patterns found in `.cvsignore' are only valid for the directory that contains them,
not for any sub-directories.

That seems to imply a per-directory view, which would explain the behaviour you are seeing. It looks as though you need a CVS directory within each build subdirectory, and a .cvsignore with, say,
* .* (or .[!.]*)
in each directory. That's not a bad idea. It leaves the directory structure intact, at the cost of having to be specific with the 'clean' target. Eclipse will be *much* happier. So will I.


J.Pietschmann wrote:
the .cvsignore features a quite prominent "build" entry. However,
Eclipse thinks there is s lot of uncommitted stuff in this directory,
which I find annoying. Isn't it supposed to ignore this?
Even more unfortunate: the commandline cvs complains about missing
CVS directories in build/ and subdirectories too on commit, which
means it somehow descends there. According to the cvs info page this
shouldn't happen either.
Am I doing something wrong?

-- Peter B. West

--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, email: [EMAIL PROTECTED]

Reply via email to