Re: DogCows or Polymorphism in the Hurd

[Jonathan S. Shapiro]

On Sat, 2006-02-04 at 12:43 +0100, Patrick Negre wrote: 
> Jonathan S. says :
> 
> > They will see whatever type was bound for the name they open. In your
> > example above, if they open foo.tar.bz, they will see a file. If a
> > translator has bound "foo" to a directory view on the same object,
> > attempts to open "foo" will see a directory.
> 
> Is it really a good idea to create new name in the filesystem tree ? I mean, 
> a 
> file must have only one path, create new nodes with binding is confusing.

Why must a file only have one path? This seems like a completely
artificial and unnecessary restriction.

> I don't understand the need of binding the "foo.tar.bz" in a directory "foo".
> Isn't it just possible to the application who need the tree view of 
> foo.tar.bz 
> to access it as a directory ?

Yes, it is possible to do that, but if a second application needs to
reference the content, you either need to hand it the descriptor or a
path to that content. If you need a path, you need to make a binding.

In the case of the Windows setup.exe example, the problem is that setup
discovers the location of the distribution directory by computing the
path to the setup.exe binary, so if you have

foo/bar/setup.exe

the setup program assumes that the rest of the installation is described
by a file

foo/bar/setup.ins

and any needed files can be found underneath foo/bar/

If the distribution was shipped as a ZIP file, then if you do NOT do a
binding there is only foo/bar.zip, so attempts to open the other files
fail.

The reason to add the binding is to allow oblivious programs to operate
naturally. It is not a requirement if all you want to do is browse.

shap