www.libssh2.org | Daily snapshots | Mailing list archive | Docs | Examples | github

Archive Index This month's Index

Subject: Re: tickets #250, #254, #256 writes -> LIBSSH2_ERROR_BAD_USE ->stuck

Re: tickets #250, #254, #256 writes -> LIBSSH2_ERROR_BAD_USE ->stuck

From: Chris Hecker <checker_at_d6.com>
Date: Tue, 19 Feb 2013 11:59:23 -0800

Or, I guess another question is, is there any way to recover a channel
from this state (where's it's returning BAD_USE from send_existing), or
is it completely toast?


On 2013-02-19 01:35, Chris Hecker wrote:
> Hi, I recently started using libssh2 called from perl's Net::SSH2, and I
> started running into this exact issue
> (https://trac.libssh2.org/ticket/256) and was wondering if yet another
> complication here might be calls from a scripting language where thunks
> are used to get into libssh2, and those might not preserve the same
> exact pointer on subsequent calls, even if the data from the scripting
> language is the same. It's almost like the failure case in this ticket
> https://trac.libssh2.org/ticket/250 due to the buffer being on the
> stack, but it's happening at a different level than the stack issue.
> This old thread from 2011 is also relevant:
> http://comments.gmane.org/gmane.network.ssh.libssh2.devel/5080
> It just seems like requiring the exact pointer on subsequent calls is
> going to be brittle no matter what, and it's going to keep breaking in a
> lot of different ways.
> I guess, practically speaking, is there a patch to fix the known cases
> of this bug, or a timeline for fixing it/them?
> Thanks,
> Chris
libssh2-devel http://cool.haxx.se/cgi-bin/mailman/listinfo/libssh2-devel
Received on 2013-02-19

the libssh2 team