You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@mina.apache.org by "Bobby Powers (JIRA)" <ji...@apache.org> on 2010/12/14 20:17:00 UTC

[jira] Created: (SSHD-101) [PATCH] [sftp] send correct amount of data in SSH_FXP_READ

[PATCH] [sftp] send correct amount of data in SSH_FXP_READ
----------------------------------------------------------

                 Key: SSHD-101
                 URL: https://issues.apache.org/jira/browse/SSHD-101
             Project: MINA SSHD
          Issue Type: Bug
            Reporter: Bobby Powers
         Attachments: 0001-sftp-send-correct-amount-of-data-in-SSH_FXP_READ.patch

Pretty sure this was just a typo, but the SFTP spec, draft 13, specifies
in section 8.2.1 that "The server MUST not respond with more data than
is specified by the 'length' parameter".  We were always responding with
0-32k's worth of data, but (at least) the Mac transmit client was asking
for only 16k chunks.  This confused it, as I believe it got the length
via stat, and then would get an early (from its point of view EOF).
Instead of closing the file and moving on, it would hang the transfer.

(see attached patch)

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


[jira] Updated: (SSHD-101) [PATCH] [sftp] send correct amount of data in SSH_FXP_READ

Posted by "Bobby Powers (JIRA)" <ji...@apache.org>.
     [ https://issues.apache.org/jira/browse/SSHD-101?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Bobby Powers updated SSHD-101:
------------------------------

    Attachment: 0001-sftp-send-correct-amount-of-data-in-SSH_FXP_READ.patch

> [PATCH] [sftp] send correct amount of data in SSH_FXP_READ
> ----------------------------------------------------------
>
>                 Key: SSHD-101
>                 URL: https://issues.apache.org/jira/browse/SSHD-101
>             Project: MINA SSHD
>          Issue Type: Bug
>            Reporter: Bobby Powers
>         Attachments: 0001-sftp-send-correct-amount-of-data-in-SSH_FXP_READ.patch
>
>
> Pretty sure this was just a typo, but the SFTP spec, draft 13, specifies
> in section 8.2.1 that "The server MUST not respond with more data than
> is specified by the 'length' parameter".  We were always responding with
> 0-32k's worth of data, but (at least) the Mac transmit client was asking
> for only 16k chunks.  This confused it, as I believe it got the length
> via stat, and then would get an early (from its point of view EOF).
> Instead of closing the file and moving on, it would hang the transfer.
> (see attached patch)

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


[jira] Resolved: (SSHD-101) [PATCH] [sftp] send correct amount of data in SSH_FXP_READ

Posted by "Guillaume Nodet (JIRA)" <ji...@apache.org>.
     [ https://issues.apache.org/jira/browse/SSHD-101?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Guillaume Nodet resolved SSHD-101.
----------------------------------

       Resolution: Fixed
    Fix Version/s: 0.6.0
         Assignee: Guillaume Nodet

Thx for reporting and patching this!

> [PATCH] [sftp] send correct amount of data in SSH_FXP_READ
> ----------------------------------------------------------
>
>                 Key: SSHD-101
>                 URL: https://issues.apache.org/jira/browse/SSHD-101
>             Project: MINA SSHD
>          Issue Type: Bug
>            Reporter: Bobby Powers
>            Assignee: Guillaume Nodet
>             Fix For: 0.6.0
>
>         Attachments: 0001-sftp-send-correct-amount-of-data-in-SSH_FXP_READ.patch
>
>
> Pretty sure this was just a typo, but the SFTP spec, draft 13, specifies
> in section 8.2.1 that "The server MUST not respond with more data than
> is specified by the 'length' parameter".  We were always responding with
> 0-32k's worth of data, but (at least) the Mac transmit client was asking
> for only 16k chunks.  This confused it, as I believe it got the length
> via stat, and then would get an early (from its point of view EOF).
> Instead of closing the file and moving on, it would hang the transfer.
> (see attached patch)

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