On Wed, Sep 19, 2007 at 04:58:49PM +0200, Oleg Verych wrote:
>
> I actually contacted Herbert back in summer. He told me, that he will
> take a look. But obviusly (including BTS's bugs) he have no time for
> dash.
Sorry, I've been travelling a lot since July so your email
is still in my backlog.
> I did rewrite of `test` built-in. Now i have 50% size deduction, but
> this is without "-o", "-a" things, that POSIX already dismissed. I
> just tried to address `!' and `(' as first parameters problem. This is
> also a very-very old problems of all shells, as far as i can see in
> historic records. Managed to do a rewrite :)
>
> I'll finish by adding "-a" and "-o" by recurcive handling later.
Sounds good.
> Permission checks currently done by stat() plus some kind of uid/gid
> checks from bash.
>
> I propose to use euidaccess(), because it have all credentials checks
> in the kernel (including SELinux, capabilities, file systems, etc.)
> I just saw BR in bash, but then the same copy/paste stuff in `test`.
As long as euidaccess does the right thing for root/execute
then this should be fine.
> In case if it is not exising in Hurd or BSD, then access() can be
> used, because `dash` is not SUID anyway.
Actually access(2) doesn't do the right thing but it should
be easy enough to use the existing code if euidaccess doesn't
exist.
Cheers,
--
Visit Openswan at http://www.openswan.org/
Email: Herbert Xu ~{PmV>HI~} <[EMAIL PROTECTED]>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
