[sane-devel] saned/Epson problems - second try

Henning Meier-Geinitz henning@meier-geinitz.de
Sun, 5 Jan 2003 02:37:49 +0100


On Sat, Jan 04, 2003 at 06:07:58PM -0300, John Coppens wrote:
> On Sat, 4 Jan 2003 15:46:00 +0100
> Henning Meier-Geinitz <henning@meier-geinitz.de> wrote:
> > 
> > For me it really looks like the old race condition bug in saned.
> > 
> Just went to the Foundation's offices to do a few tests, but things were
> rather hectic, and I couldn't get access to the scanner computer. 
> I can assure though, that no different versions are installed as you
> suspected. I just installed Linux on the machine, after formatting, and
> just in case, I ran locate saned. It is version 1.0.9... And xsane is
> version 0.89.

Then I don't know. Please provide the complete debug output of
/usr/local/sbin/saned -d128
while running xsane from another computer.

> Being a bit bored, I logged in remotely, and started 
> xsane --display:etcetc from the server (Isn't Linux great?). Anyway, it
> was not an easy way to do many tests, but I did discover something
> worthwhile, I think:
> When the 'Invalid Argument' error appears (always if not logged as root)
> when changing modes, you can make it go away by opening the preferences
> window, and closing it again (no 'apply' or anything. Just open/close).
> The mode change then works (once).

But you did scan using the net backend in this case, also? I mean just
redirecting the display of xsane to another computer doesn't change
how sane works, it's just a local scan. If the proble also occured
when using xsane without the net backend, saned/net is not involved at

> Then I could lay my hands on the (president's) Windows machine, and
> installed Xsane Win32 - version 0.90... Works a charm! No problems at all,
> and he was very impressed at being being able to use the scanner of
> another machine.

Can be a timming issue, too. We had problems before 1.0.7 that only
occurred when scanning over the net on localhost and only happend with
fast computers.