Re: emacs-w3m?

["Stephen J. Turnbull" <stephen@xemacs.org>]

Richard Stallman writes:
 >       Adding a file whose parent directory is not versioned will implicitly
 >       add the parent, and so on up to the root. This means you should never
 >       need to explicitly add a directory, they ll just get added when you
 >       add a file in the directory.
 > 
 >     The second sentence explains the rationale behind this feature.  But
 >     in this case, Bob mistakenly added a single file, and that ended up
 >     adding the whole directory.  By contrast, "cvs add" is explicitly
 >     non-recursive, AFAIR.
 > 
 > The feature is clearly a convenience in the case where you are not
 > making a mistake.  Should `bzr add' ask for confirmation before doing
 > this?

No.

The feature described above doesn't invite mistakes.  Both a careful
reading of the documentation, and experimentation with bzr, indicate
that it adds (a) the files requested on the command line, and (b) the
parent directories, and nothing else.

The implicit "add everything" command, "bzr add", is somewhat
dangerous, but that feature is really only useful when importing a new
project, ie, in the idiom "bzr init; bzr add".  To add new files to an
existing project, it's almost always preferable to use an explicit
argument, because there is very often junk like test files or personal
notes that it would pick up.  Usually either a single file name, a
single glob, or "." (meaning "everything in this subdirectory", such
as when adding a large package like CEDET to Emacs) is sufficient.
Those are hardly inconvenient.