From: James Gray <james@...>
Date: 2008-04-30T20:25:25+09:00
Subject: Re: Repeatable bug in Net::Telnet EOL translation

On Apr 30, 2008, at 5:27 AM, Brian Candler wrote:

>> I've created a patch against trunk that I believe adds this check and
>> attached it to this message.  Could you please confirm that this  
>> patch
>> resolves the regex matching issue you are seeing?
>>
>> If anyone sees any problems with this fix, please feel free to point
>> them out.
>
> Thank you for your quick response. In my test rig, your patch makes  
> the
> problem go away for operation with 'Telnetmode'=>true (the default).
>
> However the problem is still there if you set 'Telnetmode'=>false in  
> the
> client. This is because there's a separate piece of code for this  
> case:
>
>           # Not Telnetmode.
>           #
>           # We cannot use preprocess() on this data, because that
>           # method makes some Telnetmode-specific assumptions.
>           buf = c
>           buf.gsub!(/#{EOL}/no, "\n") unless @options["Binmode"]
>           rest = ''

Good catch.

> So similar logic is needed to split the string if it ends with \r. The
> following seems to work - does it look OK to you?
>
> --- telnet.rb.orig      2008-04-30 11:14:57.000000000 +0100
> +++ telnet.rb   2008-04-30 11:22:01.000000000 +0100
> @@ -570,14 +570,20 @@
>               buf = preprocess(c)
>               rest = ''
>             end
> -         else
> +          else
>            # Not Telnetmode.
>            #
>            # We cannot use preprocess() on this data, because that
>            # method makes some Telnetmode-specific assumptions.
> -           buf = c
> -           buf.gsub!(/#{EOL}/no, "\n") unless @options["Binmode"]
> +           buf = rest + c
>            rest = ''
> +           unless @options["Binmode"]
> +             if pt = buf.rindex(/\r\z/no)
> +               buf = buf[0 ... pt]
> +               rest = buf[pt .. -1]
> +             end
> +             buf.gsub!(/#{EOL}/no, "\n")
> +           end
>           end
>           @log.print(buf) if @options.has_key?("Output_log")
>           line += buf

Works for me.

I've committed our combined patch to trunk and the Ruby 1.8 branch.   
Should be all fixed up now.

James Edward Gray II