[sane-devel] 1.0.25 is out, now what?

Olaf Meeuwissen paddy-hack at member.fsf.org
Wed Oct 21 12:32:37 UTC 2015


Alessandro Zummo writes:

> On Mon, 19 Oct 2015 22:20:25 +0900
> Olaf Meeuwissen <paddy-hack at member.fsf.org> wrote:
>
>> Adding support for new devices, improving functionality for devices that
>> are already supported and fixing bugs will be part of 1.0.26 as well but
>> I think that that goes without saying.  Said it anyway and maybe that is
>> just as well.  Better be explicit.
>
>  Thanks for your hard work.

Thanks for the thanks ;-)  It's nice to know at least some people
appreciate what one's doing.

>  Regarding coolscan3, epson2 and epsonds,
>  I have no major additions.
>
>  The first has probably a lot of bugs, but I miss any detailed
>  specification from Nikon for most of the scanners.

No specs is a problem that probably affects most of the backend
developers.  Reverse engineering is no fun and even less so when you
don't have direct access to a device.  Maybe we could think about how to
approach manufacturers to get access to specs and/or devices (and under
what conditions because we still want to be able to release source
code under LGPL-like conditions!) but that is not for 1.0.26.  Perhaps a
discussion about this is better conducted off-list.  Anyway, this is not
something that affects 1.0.26 directly.

> [snip epson2 and epsonds stuff]

>  I welcome C99/C11 and the new libusb.

C11?  Too early, I think.  C99?  So far we have 3 ayes and 0 nays.

>  There's probably still some bug in the usb code. I have
>  intermittent results with some scanners on usb2
>  and with others on usb3. switching among the two usb
>  standards solves them all.

I'm curious about USB related issues that other backends are facing that
might be solved by switching to libusb-1 as the default.  I intend to
keep libusb-0 as a fallback for 1.0.26.

>  Infrared support could be finally enabled as well. my own
>  tiffscan frontend already handles it and I had some
>  good results using it with coolscan3 and doing
>  IR scratch removal with vuescan directly on the produced tiff files.

IR support?  I knew you were gonna bring that up ;-) but that's a SANE
API change.  Given the schedule for 1.0.26, I don't think that is doable
within that time frame.  If we are going to change the API, there are
probably a *lot* of things we should work out before we can commit to
anything.  One of those things would be a clearly documented mechanism
for API changes so that backends *and* frontends can deal with them in a
graceful and sane manner (pun intended) and changes can be made in a
more incremental fashion than library major so-version bumps (if at all
possible).

There's lots more but I'll keep that for later so I can focus on getting
a 1.0.26 out in time for the major Linux distribution's spring releases.
If there are other releases SANE should care about, please say so!

Hope this helps,
-- 
Olaf Meeuwissen, LPIC-2            FSF Associate Member since 2004-01-27
Support Free Software               Support the Free Software Foundation
https://my.fsf.org/donate                        https://my.fsf.org/join
 GnuPG key: F84A2DD9/B3C0 2F47 EA19 64F4 9F13  F43E B8A4 A88A F84A 2DD9



More information about the sane-devel mailing list