www.delorie.com/archives/browse.cgi   search  
Mail Archives: cygwin-developers/1998/12/03/08:28:28

From: cgf AT cygnus DOT com (Christopher Faylor)
Subject: console mode and \r, \n, etc.
3 Dec 1998 08:28:28 -0800 :
Message-ID: <19981203112448.A10350.cygnus.cygwin32.developers@cygnus.com>
Mime-Version: 1.0
To: cygwin32-developers AT cygnus DOT com
Cc: Sergey Okhapkin <sos AT prospect DOT com DOT ru>

I'm investigating Corinna's problems with console mode stuff.

What a mess.

I didn't simplify matters with my change to fhandler_console::read1 but
I see why I did things the way I did.  I was essentially trying to get
bash/ash 'read' commands to work correctly with newline endings.  It
is really tricky given the combination of binmode, ICRNL, INLCR, IGNCR,
etc.

It seems that the only way to get the pinpoint control that cygwin needs
is possibly to avoid LINE_INPUT mode or at least to avoid doing a
ReadLine on the console input.  There is an apparent bug in the
interaction between PeekConsoleInput and ReadLine that has led me to
believe that we should avoid doing a ReadLine on a console.

So my question is, does anybody object to doing "character at a time"
reads on the console?  This will slow down console reads slightly but
I'm hoping to get as much right as possible with console input.  For one
thing, I think I've identified the consistently reported problem of
cursor keys not working on Windows 95.  That should be fixable.

Sergey, did you make the change to fhandler_console to make it use
ReadLine in some cases?  If so, what was your reason for doing this?

-- 
cgf AT cygnus DOT com
http://www.cygnus.com/

- Raw text -


  webmaster     delorie software   privacy  
  Copyright © 2019   by DJ Delorie     Updated Jul 2019