Re: I discussed reading directories as files with jra, Stallman, and loic

Matthias Urlichs (smurf@noris.de)
21 Jun 1999 08:48:27 +0200

root <reiser@ceic.com> writes:
> I convinced him that directories can do this, and do it cleaner than
> that MS crud they use as an FS model. The directories need a few
> features added though.
>
> First, they need to be able to have a file that when you read them they
> resolve to.

Why not just do it this way..:

int do_open(char *name, int flags) {
int res = open(name, flags);
if(res == -1 && errno == EISDIR) {
char defname[strlen(name)+9];
sprintf(defname, "%s/default", name);
res = open(defname, flags);
}
return res;
}

> Next files need to be able to inherit stat data, so that a file can
> share its modification time with its parent directory, so that modifying
> the file changes the mod time on the directory.
>
Hmmm... why does the directory need a modification time at all, then? Just
stat the file first (reverse code than above, check for ENODIR).

> * implement a write which when done to a file with an inheritance flag
> set on it will look for and parse an rdf style syntax to determine how
> to perform the write.
>
Ugh, I do think that one should be in a library somewhere.

In the specific case of /etc/services, which should be ordered but only
because the first matching entry "wins". /etc/services is a good example
because I wonder what you'd use as file name in the first place -- the
service name? The number? Any one of the aliases? IMHO the file doesn't
really fit the directory lookup paradigm at all. So why try and force it
into one?

-- 
Matthias Urlichs  |  noris network GmbH   |   smurf@noris.de  |  ICQ: 20193661
The quote was selected randomly. Really.    |      http://www.noris.de/~smurf/

-- 
What is the approximate air speed of an unladen swallow ?


-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/