ftp://ftp.mostang.com/pub/sane/sane-0.74.lsm
ftp://ftp.mostang.com/pub/sane/sane-0.74.tar.gz
ftp://ftp.mostang.com/pub/sane/sane-0.73-0.74.diff.gz
For a list of mirrors, see http://www.mostang.com/sane/source.html
(note that it may take a couple of hours before the mirrors pick the
new version up).
The relevant section from the NEWS file is appended below. There were
_lots_ of changes---thanks to everybody who has contributed to this
release!
An important change as far as documentation is concerned is that now
every backend is being described by a short "description" file (see
backend/FOO.desc). If you're maintaining a SANE backend, please make
sure that you keep that file up to date. That way, we can ensure that
the web page (http://www.mostang.com/sane/sane-backends.html) always
contains the latest list of supported devices.
Cheers,
--david
---------------------------
New with 0.74:
* GIMP auto-detection should work again.
* Service name for remote scanning changed from "saned" to "sane".
Be sure to update your /etc/services and /etc/inetd.conf if you
use remote scanning! We generally try to avoid such incompatible
changes, but the name "saned" was definitely wrong, so it seemed
better to change it now.
* Thanks to Matto Marjanovic work, each backend now comes with a
.desc file that describes exactly what devices it supports. These
description files can be automatically translated into various
other formats, such as HTML. See:
http://www.mostang.com/sane/sane-backends.html
for an example as to what this can do for you.
* New backend for Kodak DC25 digital cameras (by Peter Fales).
* Updated Artec backend (by Chris Pinkham).
* Updated Microtek backend (by Matthew Marjanovic)
o Complete rewrite of bit-shuffling, buffering, and color-handling code.
o Improved rescaling algorithm for expanded resolution scans.
o Support for 600GS (and maybe ZS, too) (thanks to Oliver Neukum).
o Support for document autofeeder and IIG (thanks to Ross Crawford).
o Fixed sane_cancel.
o sane_get_parameters size estimates are now accurate to the last bit.
o get_scan_status timeout increased (for 600GS), status code fixed.
o Fixed parsing of 3-pass bit in INQUIRY data.
o Stopped sending gamma tables to scanners that don't want them.
o Made precalibration a bit more robust (always attempt at least one
scan line now).
o Much, much code clean-up.
o Tested & working with saned. (Atrocious hack so sane_read
handles buffers smaller than one scanline.)
o Auto-expand pre/post hold buffers as necessary (fixes some problems
with single-pass color scans).
o Added configuration file option to disable precalibration logic.
o Fixed document size calculations.
o Added more informative scsi-II note.
o Remove misnomer "Highscan" from manpage.
o Updated man-page.
* Updated Microtek2 backend (by Bernd Schroeder)
o changed the code for gamma calculation and added a custom gamma table
to the options. In some cases this requires an additional memcpy(),
but the code is now leaner and cleaner.
o fixed the bugs that the backend didn't compile with non gcc compilers.
o added an option to control exposure times.
o rewrote the code that processes the data that is received from the
scanner and transfers it into the frontend buffer. The smallest unit of
data that can be copied into the frontend buffer is now a pixel, no
longer a complete line.
o added (a yet) undocumented option "strip-height" that allows to control
the number of lines that is read from the scanner with one "read image".
o fixed a bug that caused scanimage to sigsegv when calling sane_exit()
without having written all data into the output file.
o added code to support scanners with more than 8 bit output per color.
Due to the lack of such a scanner this functionality is COMPLETELY UNTESTED
and there are some potential sources of bugs, but at least one could give
it try.
o added sanei_config_attach_matching_devices()
o improved the code for the check if it really is Microtek SCSI-II scanner
o fixed the "lineart inverted" bug
o The threshold option in lineart mode is now in the enhancement group
o changed the default for the resolution
o The values for the SANE_Params struct are calculated more precisely
o dito the number of scan lines that fit into the frontend buffer
o changed some return stati ( replaced SANE_STATUS_INVAL with
SANE_STATUS_IO_ERROR where the first one is not allowed)
o completely rewrote the end of scan and cancel scan handling
o fixed another bug that caused xscanimage to crash with sigsegv
under some circumstances (check of the inquiry result was wrong)
o added model code for a Vobis Highscan
o support for new format of configuration file (including
"option <opt> <val>")
* Updated Nikon Coolscan backend (by Didier Carlier).
* Updated UMAX backend (by Oliver Rauch):
o Cancelling a scan now works, cancelling a preview still makes problems!
o Preview fix is activated for Astra 600S, Astra 610S and Vista S6E
scanners. (calibration by driver still does not work for these scanners).
- removed button for preview patch
o Quality calibration / calibration by driver now works for same scanners.
- added selection for type of shading data calculation
o Solved problem with black area at top of the image
(happend only with some SCSI-adapters).
o Added gamma download format type 0 and type 1 for older scanners.
o Added Astra 1220S as supported scanner:
- added 36 gamma input bits support for Astra 1220S
- added 36 output bits support, but there is no frontend that can
handle it!
o Added inquiry for Escom Image Scanner 256 (UMAX UG80).
* OS/2 (by Yuri Dario):
o Updated SCSI support so sanei_scsi_find_devices is supported now.
o Epson parallel support added (by Rocco Caputo <troc@netrus.net>)
-- Source code, list archive, and docs: http://www.mostang.com/sane/ To unsubscribe: echo unsubscribe sane-devel | mail majordomo@mostang.com