> E.g., consider the netproxy on machine B in Figure 3.1 of the
> standard. All it sees is a list of device name strings. It looks to
> me like this needs some thought.
Yes ... anything better than my simple approach would be appreciated.
> Andy> case, because : a) It is not possible to exchange the
> Andy> authentication data via standard SANE calls yet. (This is a
> Andy> deficiency we should try to alleviate somehow. The easiest
> Andy> way would be to read .netsanerc from the client, as it would
> Andy> not cause changes in SANE. Another would be to add a sane_auth
> Andy> call ...)
> Agreed. This needs to be addressed, eventually.
Yes, because we need some standard method of requesting authentication
data ...
Requesting the auth itself is easy : Simply return E_AUTH on any SANE
call which needs more permissions that the client currently has.
> Andy> c) host1 can
> Andy> connect directly to host3, if desired. (I do not think we need
> Andy> proxy-sane ... do we ;-) ?
> I actually like the idea (see, I like to work around firewalls...).
Yes ... nasty things, if you do not have transparent proxying like
in the newer Linux kernels ...
> But it's not something we should spend extra effort on. But if it
> falls off naturally, I'd call it a feature.
Yep.
For now I will probably simply do nothing about it, to make it work
ASAP.
CU,ANdy
-- ============================================================================== Andreas Beck | Email : <becka@sunserver1.rz.uni-duesseldorf.de>
--- Source code, list archive, and docs: http://www.azstarnet.com/~axplinux/sane/ To unsubscribe: echo unsubscribe sane-devel | mail majordomo@azstarnet.com