From: cgf AT cygnus DOT com (Christopher Faylor)
Subject: Re: B20 patch: CR handling in fhandler_console
2 Dec 1998 19:32:19 -0800
Message-ID: <19981202220732.A6801.cygnus.cygwin32.developers@cygnus.com>
References: <3665D672 DOT A1722461 DOT cygnus DOT cygwin32 DOT developers AT cityweb DOT de>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
To: DJ Delorie <dj AT delorie DOT com>, cygwin32-developers AT cygnus DOT com

On Wed, Dec 02, 1998 at 09:42:42PM -0500, DJ Delorie wrote:
>Corinna Vinschen wrote:
>> in fhandler_console::read1() a short piece of code
>> ignores CR if it is the one and only char in console input.
>> 
>> I have found no circumstance, in which this code makes sense.
>
>Be careful.  If the console is in text mode, the \r character
>will be stripped (i.e. crlf->nl conversion).  If the user
>asked to read one character, and it is stripped, the function
>will report that zero characters were read.  This is an EOF
>indicator, even though it is not really EOF.
>
>What the code should do is *if* the \r would be stripped,
>it should read the following \n.  If it can tell that
>the \r is alone (no \n) but would be stripped, it should
>probably change it to a \n.  If the console is in binary
>mode (no stripping), it can safely return a \r character.

At that point in the function, the console is *guaranteed* to be in cooked
mode.  If it sees a \r, then it should be ignored.

I think the code is correct, but then I would, because I wrote it.

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