[sane-devel] Re: updated Fujitsu SANE backend

[Charles Duffy]

m. allan noah <anoah <at> pfeiffer.edu> writes:

> upright fujitsu- turn it over, start from the bottom. mostly screws,
> rarely tabs, never glue.

Took off the four obvious screws on the bottom, didn't know where to
go from there. (It's sitting upside-down in the office presently --
I'm at home, but if I come upon information that will substantially
improve its survivability, I'll certainly run back).

> > Also, I note that "startbutton" hasn't been replaced with simply
> > "button", as needed to work with the button-daemon.
> 
> yeah- thats pretty far down on the list. besides, i think extending
> buttond to understand that options button_* are also buttons might
> be better. then we could have button_scan and button_sendto...

Hmm. The current implementation allows numbered buttons (by putting a
different value in the button variable), but that's certainly not as
good as named ones.  However, if I understood the relevant SANE2 spec
portion correctly, the current plan is to use a bitfield for querying
button states -- so relying on named buttons may not be a good
long-term solution (unless the spec is extended to allow a list of
names associated with each bit to be provided to the frontend).

I also wouldn't want to eliminate support for numbered buttons, if for
no other reason because I don't know the button label <-> number
mappings for all the relevant scanners... but if nobody objects to
adding named button support (for button_* options) in addition to
numbered button support, I'm game to try throwing together a patch...
if, of course, I still have hardware to test on come Monday.