Mailing-List: contact cygwin-apps-help AT sourceware DOT cygnus DOT com; run by ezmlm Sender: cygwin-apps-owner AT sourceware DOT cygnus DOT com List-Subscribe: List-Archive: List-Post: List-Help: , Delivered-To: mailing list cygwin-apps AT sources DOT redhat DOT com Date: Wed, 2 May 2001 10:31:40 -0400 From: Christopher Faylor To: cygwin-apps AT cygwin DOT com Subject: Re: Forcing SYSTEMROOT (opinions needed) Message-ID: <20010502103140.D28400@redhat.com> Reply-To: cygwin-apps AT cygwin DOT com Mail-Followup-To: cygwin-apps AT cygwin DOT com References: <20010501232846 DOT C26438 AT redhat DOT com> <3AF00F4D DOT B3F70233 AT yahoo DOT com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.3.11i In-Reply-To: <3AF00F4D.B3F70233@yahoo.com>; from earnie_boyd@yahoo.com on Wed, May 02, 2001 at 09:44:45AM -0400 On Wed, May 02, 2001 at 09:44:45AM -0400, Earnie Boyd wrote: >Christopher Faylor wrote: >> >> On Wed, May 02, 2001 at 01:14:39PM +1000, Robert Collins wrote: >> > >> >> -----Original Message----- >> >> From: Christopher Faylor [mailto:cgf AT redhat DOT com] >> >> Sent: Wednesday, May 02, 2001 1:21 PM >> >> To: Christopher Faylor >> >> Subject: Re: Forcing SYSTEMROOT (opinions needed) >> >> >> >> >> >> So we have to trade the possibility of someone wanting >> >> complete control of >> >> his environment versus the possibility of someone not >> >> specifying SYSTEMROOT >> >> but needing it for the program that is about to be run. >> > >> >This ones easy. Remove winsock/socket support from cygwin. See the >> >problems gone :] >> >> Yeah, but then people would claim that Cygwin was "unstable" and "always >> breaking network support". >> >> >>Should I flip a coin? >> > >> >I'd add systemroot. If it's needed from cygwin running, not having it >> >is like trying to run hurd without a microkernel - if you want it >> >different you have to fiddle yourself! >> > >> >We can always do a switch to control it down the track. >> >> I guess we could add a cygwin_keep_environment_clean () call or something. >> Is there a standard way to control this type of behavior on UNIX. I'm thinking >> that there is something like sysconf or ulimit or something but more generic >> than that. I'm drawing a blank, though. >> > >The other option is to leave it alone and document it. We've discovered >the problem how many times in the past how many years? I guess my memory is failing because I don't recall this problem. cgf