[FreeNX-kNX] nxagent session gets lost, user gets new session even though one already exists

[Mario Becroft]

Marcelo Boveto Shima <marceloshima at gmail.com> writes:

> Sorry, this diff don´t help at all.
> This is a bug at slave mode.
> Set ENABLE_SLAVE_MODE="0" and all will be fine.

Thanks for your help with this. Disabling slave mode does seem to
prevent the specific issue I was having, but it is still broken
somehow. With slave mode disabled, there is a slightly different issue
where nxserver never notices that the session is suspended. So once you
have created a session, it always remains in running state even after
you suspend it.

This does not prevent you from suspending and resuming the session, but
at least it causes incorrect status to be reported in nxserver --list.

The problem looks slightly similar to the one I was having before. In
this case, when you kill the client nxssh, this happens:

--8<---------------cut here---------------start------------->8---
NX> 596 Error: Session  failed. Reason was: Session: Display failure detected at 'Sun Jan 25 10:22:04 2009'.
(/usr/NX/bin/nxserver)(18933)(Sun Jan 25 10:22:04 NZDT 2009) AAAA cmd=(NX> 596 Error: Session  failed. Reason was: Session: Display failure detected at 'Sun Jan 25 10:22:04 2009'.)
NNNN(/usr/NX/bin/nxnode)(19338)(Sun Jan 25 10:22:04 NZDT 2009) saying that session status is suspending
NX> 1009 Session status: suspending
NNNN(/usr/NX/bin/nxnode)(19338)(Sun Jan 25 10:22:04 NZDT 2009) saying that session status is suspended
NX> 1005 Session status: suspended
(/usr/NX/bin/nxserver)(18933)(Sun Jan 25 10:22:04 NZDT 2009) server_nxnode_echo: NX> 596 Error: Session failed. Reason was: Session: Display failure detected at 'Sun Jan 25 10:22:04 2009'.
--8<---------------cut here---------------end--------------->8---

It gets the "NX> 596 Error..." command, but after that it no longer
processes any commands, such as the "NX> 1009" and "NX> 1005"
commands. Even after restoring and suspending the session again, it
still does not process any commands.

The main reason why I have not tried disabling slave mode before is
because it is claimed to be slower. However, having test it just now it
seems the same speed, or possibly faster! Is this expected?

BTW, what issue does your earlier patch fix? Should I apply it as well?

-- 
Mario Becroft <mb at gem.win.co.nz>