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: <mailto:cygwin-apps-subscribe AT sources DOT redhat DOT com>
List-Archive: <http://sources.redhat.com/ml/cygwin-apps/>
List-Post: <mailto:cygwin-apps AT sources DOT redhat DOT com>
List-Help: <mailto:cygwin-apps-help AT sources DOT redhat DOT com>, <http://sources.redhat.com/lists.html#faqs>
Delivered-To: mailing list cygwin-apps AT sources DOT redhat DOT com
Message-ID: <20010311042451.64969.qmail@web5410.mail.yahoo.com>
Date: Sat, 10 Mar 2001 20:24:51 -0800 (PST)
From: Steve O <simgenome AT yahoo DOT com>
Reply-To: bub AT io DOT com
Subject: rxvt and getpass()
To: cygwin-apps AT sourceware DOT cygnus DOT com
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii

Hi and thanks for letting me post to cygwin-apps,

At some point I'd like to see if I could include the 
No-X rxvt port into the cygwin distro.  Before that 
happens, though, I'd like to fix a bug that's been
challenging me for some time. 

The tip of the iceberg is that getpass() opens /dev/tty
in order to print the password prompt and to get the
user's response.  Opening /dev/tty is equivalent to 
getting the handles of the win32 console associated with
the process (as far as I can tell).  When using rxvt, 
the console is hidden, and so there's no way for the 
user to respond to, or even see, the password prompt.

The first order solution is to edit out the code in getpass
that opens /dev/tty, but that's probably not the right
thing to do.  Instead, it seems that I should be doing
some extra step to let cygwin know that /dev/tty should
refer to my tty/pty pair.   Is there a way I can make the
open("/dev/tty") fail, perhaps with a chmod?  Or dup the
fds of my tty into the console handles?  I can experiment 
with different hacks, but I guess I'm hoping that there's
some obvious approach that I'm unaware of.

Ok, maybe the solution is for rxvt not to have a console
at all.  I really like this solution, but in the past
this has led to spurious consoles being created as
processes fork.  Is this a feature of windows that
no (non-gui) process shall be without console?  Or are
there some assumptions buried in cygwin that there will
always be a console?  Or am I just behind the times?

Any help is appreciated.  

Thanks,
-steve o

http://www.io.com/~bub/rxvt


__________________________________________________
Do You Yahoo!?
Yahoo! Auctions - Buy the things you want at great prices.
http://auctions.yahoo.com/