Message-Id: <199902242108.VAA15328@out5.ibm.net>
From: "Mark E." <snowball3 AT usa DOT net>
To: djgpp-workers AT delorie DOT com
Date: Wed, 24 Feb 1999 14:25:09 -0500
MIME-Version: 1.0
Content-type: text/plain; charset=US-ASCII
Content-transfer-encoding: 7BIT
Subject: Re: patches for chroot
References: <199902230504 DOT FAA82404 AT out1 DOT ibm DOT net>
In-reply-to: <Pine.SUN.3.91.990223105500.13846H-100000@is>
X-mailer: Pegasus Mail for Win32 (v3.01d)
Reply-To: djgpp-workers AT delorie DOT com

> > chroot ("c:/djgpp")
> > getcwd would then return '/bin' if the cwd is 'c:/djgpp/bin'.
> 
> Yes.
> 
> I think _fixpath should do that as well, to prevent confusion, and also to
> be compatible with Unix description that every call to chroot interprets
> its argument *relative to the previous root*.
> 

I've run into a problem with stat() and _fixpath.

void foo()
{
  struct stat finfo;

  chroot ("c:/djgpp");
  if (stat ("c:/djgpp/bin", &finfo) < 0)
    return -1;
}

The call to stat() fails because "c:/djgpp" is stripped so all it gets back 
is "/bin" and since a call to _put_path is inappropriate, the root 
expansion doesn't take place. There's also the case of chroot 
("c:/djgpp"); stat ("/bin", &finfo); that fails for a similiar reason.

One solution would be to have a flag for _fixpath to tell it when not to 
delete the root path from the output. I suppose another would be to not 
delete the root path from the output at all, but it was said above we 
want _fixpath to do that. Thoughts?

Mark

BTW, Thanks Nate for running the chroot test.

--- 
Mark Elbrecht
snowball3 AT usa DOT net http://members.xoom.com/snowball3/