commons-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Sebb (JIRA)" <j...@apache.org>
Subject [jira] Commented: (NET-335) TelnetOutputStream incorrectly doubles-up IAC when sending IAC+EOR in BINARY transmission mode
Date Fri, 15 Oct 2010 11:16:33 GMT

    [ https://issues.apache.org/jira/browse/NET-335?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12921309#action_12921309
] 

Sebb commented on NET-335:
--------------------------

The problem with option 1 is that this breaks compatibility for existing users, who expect
IAC characters to be doubled. 
Indeed if your data includes IAC not followed by EOR, it would not be sent correctly.

Note that a work-round is to use the protected _output_ field from SocketClient, which is
not subject to any filtering.

> TelnetOutputStream incorrectly doubles-up IAC when sending IAC+EOR in BINARY transmission
mode
> ----------------------------------------------------------------------------------------------
>
>                 Key: NET-335
>                 URL: https://issues.apache.org/jira/browse/NET-335
>             Project: Commons Net
>          Issue Type: Bug
>          Components: Telnet
>    Affects Versions: 1.4, 2.0
>         Environment: Windows  XP SP3, Java 1.6.0_21
>            Reporter: Chris Jones
>
> This is related to NET-180 which I reported a couple of years back, and is reported fixed
in 2.1. I've since re-visited the code, and found another issue
> background
> ----------------
> I have a TN5250 client (see rfc1205) which extends TelnetClient. Basically, the client
negotiates BINARY transmission and operates in block-mode. The client sends IAC+EOR to notify
the host that the data is complete, but the TelnetOutputStream doubles-up the IAC automatically
as if it was a data byte, which iis not the case, resulting in IAC+IAC+EOR.
> Two possible solutions come to mind:
> 1. Do not double-up IAC when in BINARY transmission in TelnetOutputStream (i.e. leave
it to the client)
> 2. Add a method to TelnetClient like -- sendCommand(byte b) {........} which would act
similar to sendAYT, only allowing the client to specify the value. 
> On balance, I like the second option, as it would give more future possibilities.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


Mime
View raw message