Re: Strange wording about io_t->io_map method

[Pierre THIERRY]

Scribit Tom Bachmann dies 25/04/2006 hora 14:59:
> This also sounds strange (to my ears), as caps are not read only or
> rw.

The cap itself is neither ro or rw. But it designates an object who
implements ro or rw methods. Or it just designates the ro or rw methods
of an rw object.

> This information is stored in the object, isn't it?

Not necessarily.

You could have these three capability types:

io_read_t
- io_read()
- io_map() [return a weak PATT]
- io_stat()

io_write_t
- io_write()
- io_map() [return an opaque PATT?]
- io_set_lenght()
- io_set_mtime()

io_readwrite_t
- io_read()
- io_write()
- io_map() [return an opaque PATT]
- io_stat()
- io_set_lenght()
- io_set_mtime()

One of each of these three types could designate the same I/O object.
But they don't give access to the same set of methods...

Virtually,
Nowhere man
-- 
address@hidden
OpenPGP 0xD9D50D8A

signature.asc
Description: Digital signature