Re: Dropbear 0.49pre build error and a ssh glitch

2007-02-01 Thread Hans-Christian Egtvedt
Matt Johnston wrote:
 On Wed, Jan 31, 2007 at 09:23:32AM -0800, William Trenker wrote:
 Thanks, Matt.  I got around it by building a standalone dbclient from
 the 0.49pre source and keeping 0.48.1 multi for everything else.  May
 I ask what your roadmap is for 0.49?  Are you weeks away, or months?
 Do you plan to run a beta?
 
 I'll probably get 0.49 out in the next few weeks. There are
 a few changes that could do with some testing, so I'll put a
 test tarball up earlier.

Can (have?) you update the config.sub and config.guess (etc..) scripts
as well? It would be great so that the new architectures added last year
will work out of the box with dropbear.

-- 
With kind regards,
Med vennlig hilsen,

Hans-Christian Egtvedt, siv.ing. (M.Sc.)
Applications Engineer - AVR Applications Lab
Atmel Norway


Re: Dropbear 0.49pre build error and a ssh glitch

2007-01-31 Thread Matt Johnston
On Mon, Jan 29, 2007 at 08:37:41PM -0800, William Trenker wrote:
 Hi Matt,
 
 I've been working with the latest development version in MTN.  Specifically
 I'm building from:
 http://viewmtn.angrygoats.net/revision/tar/d26d5eb2837f46b56a33fb0e7573aa0201abd4d5
 
 There is a small Makefile problem reported by make as a missing endif.  I
 think the problem is here:
 libtommath/Makefile.in
 ifndef IGNORE_SPEED  --- should this be commented out?

Ah yes. I fixed this locally but neglected to push the
change - should be on the server now.

 Also, I'm experiencing a problem when I try to ssh into the
 dropear-0.49server from another box such as:
 ssh [EMAIL PROTECTED] ls
 The command works fine, but doesn't terminate after the ls is finished
 listing the files.  Instead it just sits there until I hit the enter key.
 If I do the same ssh into the same box running the dropbear 0.48.1 server
 there is no problem.

This is a known problem. I modified the channel close
code to try and resolve some other issues, but didn't fix
everything. For the time being I might just revert to the
0.48 behaviour - the experimentation should've been on its
own branch in the first place.

Cheers,
Matt