Malformed Address and Private IP issue

[John Van Essen]

On Thu, 9 Mar 2006, <foner-rsync at media.mit.edu> wrote:
>     Date: Wed, 8 Mar 2006 17:15:36 -0800
>     From: Wayne Davison <wayned at samba.org>
> 
>     On Wed, Mar 08, 2006 at 01:48:37PM -0800, Jonathan Chen wrote:
>     > 2006/03/08 11:25:12 [16976] malformed address localhost.localdomain
> 
>     That can't be 2.6.6 because 2.6.6 doesn't have an error message of that
>     format.  In 2.6.6, the old "malformed address" error now outputs as
>     "error matching address" and includes the gai_strerror() text after a
>     colon.  A new error in 2.6.6 that does include the string "malformed
>     address" would have also included the gai_strerror() text after a colon.
>     Thus, that's still the old rsync running.  Perhaps it didn't really get
>     stopped?  Or perhaps it is running via inetd?
> 
> Given how often rsync versions change and how much functionality goes
> into each new one (yay!), I wonder if it might not be such a bad idea
> to have the rsync version embedded in every error message?  With most
> programs, it's likely that the user knows at least something about the
> version they're running, but since rsync is almost always run with one
> of the instantiations remote, it might make debugging easier if the
> message was explicit...

The daemon already writes its version number into the rsyncd log like so:

2006/03/07 00:00:19 [16877] rsyncd version 2.6.3 starting, listening on port 873

so Jonathan can check the log to verify the version of the daemon.

As for non-daemon errors, I don't know about embedding the version number
in every error message, but it *would* be handy if rsync knows when it has
generated an error message, then at the end of the session it could display
the version and negotiated protocol number for that session to encourage
people to include that info in any problem reports.  It has to explicitly
say if it's the receiver or the sender when it reports the version.

    John