path: root/doc/umax/umax.CHANGES

CHANGES FOR UMAX BACKEND FOR SANE
---------------------------------

umax -> sane-0.6:
-----------------

    * umax-sane-backend created:
      - umax.c split in:
        umax-cli.c, umax-scsi.c, umax-driver.c, umax-text.c
      - umax.h split in:
        umax-scanner.h, umax-scsi.h, umax-scsidef.h, umax-struct.h
      - outputs done by DBG(level,msg,args...) defined in debug.h
      - scsi-calls changed to sane-format (of sane-0.57)
      - created umax.c and umax.h for sane-backend

    * changed search-path for scsi.h and sg.h, now needs /usr/include/scsi

    * removed mistake in bound-function

    * added "Vista S-6E" as supported

    * added PowerLook-II (PL-II) as supported

    * added Astra 600S as supported


sane-0.6 / 0.61 / 0.62 -> 0.63
------------------------------

    * added exception to gpl-licence:
    >   This program is free software; you can redistribute it and/or
    >   modify it under the terms of the GNU General Public License as
    >   published by the Free Software Foundation; either version 2 of the
    >   License, or (at your option) any later version.
    >
    >   This program is distributed in the hope that it will be useful, but
    >   WITHOUT ANY WARRANTY; without even the implied warranty of
    >   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
    >   General Public License for more details.
    >
    >   You should have received a copy of the GNU General Public License
    >   along with this program; if not, write to the Free Software
    >   Foundation, Inc., 59 Temple Place - Suite 330, Boston,
    >   MA 02111-1307, USA.
    >
    >   As a special exception, the authors of SANE give permission for
    >   additional uses of the libraries contained in this release of SANE.
    >
    >   The exception is that, if you link a SANE library with other files
    >   to produce an executable, this does not by itself cause the
    >   resulting executable to be covered by the GNU General Public
    >   License.  Your use of that executable is in no way restricted on
    >   account of linking the SANE library code into it.
    >
    >   This exception does not, however, invalidate any other reasons why
    >   the executable file might be covered by the GNU General Public
    >   License.
    >
    >   If you submit changes to SANE to the maintainers to be included in
    >   a subsequent release, you agree by submitting the changes that
    >   those changes may be distributed with this exception intact.
    >
    >   If you write modifications of your own for SANE, it is your choice
    >   whether to permit this exception to apply to your modifications.
    >   If you do not wish that, delete this exception notice.

    * split resolution in x-res and y-res, flag bind-resolution

    * only sane_... functions will be exported from umax-backend,
      all umax_... functions are made static

    * removed umax-emergency-routine for SANE

    * made program-flow (esp. with do_canel) SANE-specific

    * fixed missing 3-pass parts

    * made scanmodes, sources, analog gamma, gamma download, negative,
      quality calibration, highlight, shadow only selectable if they are
      available

    * calibration-mode selectable in advanced options if available

    * warmup selectable if available

    * changed umax-scsi.c like described from Gordon Matzigkeit
      in his diff/patch-file from 21.06.97 so that the umax-backend
      should compile on non-linux-systems now - thanks!

    * split highlight and shadow in R, G and B for color-mode.

    * reset adf and uta flag when not used (in sane_start)

    * added possibility to change calibration-mode if available

    * added send_halftone_pattern to umax-scsi.c

    * changed umax-scsi.c like described from Gordon Matzigkeit
      in his diff/patch-file from 21.06.97 so that the umax-backend
      should compile on non-linux-systems now - thanks

    * added support for each color (R,G,B) for highlight and shadow

    * added "Astra 1200S" and "UC840" as supported

    * added bad bugfix for preview-problem with S6E
      (in umax-scsi.c function start_scan)

    * moved resolution-check from check_values to bound_scanarea,
      now no floatingpoint-exception will occur if no resolution is given!

    * changed range-check for width and height

    * removed bug in set_window for three-pass-rgb-mode
      -selected wrong buffer for some green and blue options
      -scanmode now is greyscale instead of rgb

    * changed start_scan(us) to better use with wid-list
      and added missing 3 wid-bytes in scan.cmd

    * removed parameter color in three_pass_function_headers,
      the color is taken from us->three_pass_color now.


sane-0.63/0.64/0.65 -> 0.66
---------------------------

    * changed gamma-data-upload for 3-pass-color-scan

    * corrected wrong calculation for highlight and shadow
      selection for rgb-values

    * changed gamma data to use 8 bit (256 bytes) or 10 bit (1024 bytes)

    * included 10/30 bits per pixel mode

    * changed bug in sane_exit (free()) that sometimes made a
      segmentation-fault

    * made patches runtime-selectable

    * changed set_window_command to support 30 bits per pixel

    * changed send_gamma_data to support three_pass_gamma and 10/30 bpp

    * changed row_len in umax_bound_scanarea to 10/30 bit mode

    * added halftone-support for ug630, uc630 and uc840,
      but I don't know if all values are ok!

    * changed CBHS-50-range to 78-178 (from 103-153)

    * changed rangecheck for highlight and shadow

    * fixed get_cbhs_range-bug in umax-scsidef.h

    * added ADF-inquiry-bits to umax-scsidef.h

    * added color_line_arrangement_mode-inquiry-byte to umax-scsidef.h

    * added CCD_line_distance-inquiry-byte to umax-scsidef.h

    * added all most unused bits and bytes in inquiry to umax-scsidef.h

    * made patches runtime-selectable


sane-0.66 -> 0.67
-----------------

    * removed bug in sane_start: ADF and UTA flags have not been set correct

    * removed bug in init_options: value range for 10 gamma-input-bits corrected

    * made custom gamma available for lineart and halftone for nonlinear effects

    * added Vista-T630 as supported scanner

    * removed test_warmup in wait_scanner because adaptec 1542 has problems
      with request_sense-command in test_warmup
      it was only used for printing a message, so it was not important.

    * if a scanner has a full inquiry-block but is not known to be supported,
      the driver will print a warning and continue instead of aborting,
      option --continue (on umax-cli) will suppress the warning!

    * changed ugly definition
      char gamma[x] -> char *gamma; gamma=alloca(x);

    * removed bug (wrong length for 10 bit-gamma) in send_gamma_data

    * changed all buffer-definitions to unsigned char and inserted
      type conversions, so there are no more warnings because of
      typemismatches.

    * corrected bad definition of set_DCF2_gamma_output_bits


sane-0.67 -> 0.68
-----------------

    * changed open config to fp = sanei_config_open(UMAX_CONFIG_FILE)

    * changed all set OPT...BIND commands, endless-loop happened in sane-0.67
      =>ane-0.67-umax-fix.diff.gz
      after that patch changed same for ten_bit_mode

    * changed some strdup() to (SANE_Char *)strdup()

    * added UMAX Astra 610S as supported scanner. it has the same problems
      in color-mode like the Astra 600S.


sane-0.68 -> 0.69
-----------------

    * changed maximum y-resolution from optical_res to y_res

    * changed sane_cancel (waitpid) like suggested by David Mosberger

    * added #include <lalloca.h> for AIX-support

    * replaced call to umax_wait_scanner with umax_do_calibration
      after umax_start_scan in umax.c

    * moved sense_handler to umax-scsi.c

    * removed call to umax_get_data_buffer_status in reader_process

    * added umax_send_shading_data, umax_send_gain_data
      and   umax_read_shading_data, umax_read_gain_data

    * changed name: umax_read_data_block -> umax_read_image_data

    * changed length offset in do_inquiry from 5 to 4
      and changed set_inquiry_length-offset from 5 to 4

    * added calib_lines to structure US and added it to
      set_window - command in umax-scsi.c

    * changed in umax-scsi.c routine identify_scanner:
      vendor and product is tested for each model instead of
      vendor=="UMAX " and only model tested!
      Because of that the format of scanner_str has changed
      from product,product,...
      to vendor,product, vendor,product,...

    * added support for Linotype Hell Jade (LinoHell Office)

    * added empty routine umax_do_calibration in umax-scsi.c, it shall be called
      after umax_start_scan!

    * replaced all umax_*scsi with sanei_*scsi, so the umax-driver
      now directly calls the sanei_scsi-commands.

    * moved sense-handler from umax.c to umax-scsi.c
      and added detection for driver-calibration-request
      sense_handler got a new parameter (void *arg).

    * changed wait_scanner for SUN: usleep(0.1sec) -> sleep(1s)
      this is only a bad bugfix that will be replaced one day!


sane-0.69 -> 0.70
-----------------

    * moved reader_process to umax-scsi.c

    * changed include <sane/config.h> for _AIX and non-_AIX

    * added missing "," in definition line for UMAX Astra 1200S in umax-scanner.h

    * added in umax-scsi.c in function identify_scanner a test that only
      warnings and errors are printed if vendor="UMAX "

    * added routine umax_order_line_to_pixel and umax_order_line

    * moved reader_process from umax.c to umax-scsi.c
      and added routine for color ordering.

    * reset length offset in do_inquiry from 4 to 5
      and set_inquiry_length-offset from 4 to 5
      because umax-S6E didn't work with it!

    * corrected error in identify-scanner in vendor-test

    * changed end-detection of list of known scanners from
      #define known_scanners 15 to last item "END_OF_LIST"

    * added routine that corrects wrong inquiry-data from
      Astra 600S and Astra 610S

    * added test and selection for do_color_ordering in check_values

    * changed __sun-code in umax_wait_scanner because the scsi-driver
      answers with IOERROR instead of DEVICE_BUSY
      (this is only a bugfix for a little time)


sane-0.70/0.71 -> 0.72
----------------------

    * inactivated selection of quality calibration (until it is supported
      by the driver)

    * changed maximum x_resolution from optical_res to max_x_res

    * reordered DBG-levels and defined them by symbols

    * changed all sizes that used type int to type size_t

    * changed wait_scanner for testing !sane_status_good,
      because testing sane_status_device_busy does not work with
      all scsi-drivers!

    * changed ccd-order in umax_order_line, the order is now
      specified by 9 values (before by 3 values), changed
      order tables and added a table for Astra 600S and
      another table for Astra 610S

    * changed ccd-order, now based on optical_resolution
      and added new routine umax_forget_line

    * added inquiry-block for UC1200S: NOT TESTED!

    * added inquiry-block for UC1260: NOT TESTED!

    * included print_inquiry into umax-scsi.c

    * added correction-routine for x_res > 600 dpi, phys_width is
      corrected in umax-scsi.c: bound_scanarea

    * changed umax-scsi: sense_handler:
      changed test for valid sense-block
      added test for button pressed

    * added output of sense-code and error-code in sense_handler

    * added options to select calibratin and scan exposure times

    * added options to select calibratin and scan lamp density

    * moved all option-names/titles/descr. to saneopts.h

    * combined bind for highlight, shadow, analog-gamma and new
      exposure time functions to rgb-bind

    * added set_WD_line_arrangement(buffer_r, WD_line_arrengement_by_driver);
      if driver is doing color-ordering from line to pixel

    * included umax-scsi.c into umax.c

    * included umax-struct.h into umax.h

    * included US-structure into Umax_Device-structure

    * added free(scanner->device->buffer) to sane_close

    * added free(scanner->device->pixelbuffer) to sane_cancel

    * added free(scanner->device->pixelbuffer) to do_cancel

    * added SANE_UNIT_MICROSECOND to exposure-times

    * added output of status-string in wait_scanner

    * moved some options for Scan-mode-block to Enhancemant-block

sane-0.72 -> 0.73
-----------------

    * disabled RGB_BIND if no such value is available

    * included readproc

    * included bound_scanarea into check_values, problem
      with wrong scanarea for UTA should be solved now!

    * all scanmode- and scansource-strings are #DEFINEd now

sane-0.73 -> 0.74
-----------------

    * added support for 12/36 bits per pixel (for Astra 1220S)

    * added support for Astra 1220S

    * included inquiry data for UG80 (Escom Image Scanner 256)

    * moved WAIT_SCANNER from umax_read_data to umax_read_image_data,
      umax_read_shading_data now works!

    * added umax_start_scan at the end of calibration by driver,
      the scanner does not hang any more if quality calibration is enabled!
      There must'nt be a "wait_scanner" in calibration by driver routine!!!

    * added usage of driver internal values for calibration by driver if
      the scanner does not give back the information about the shading data:
      width = dev->width * dev->inquiry_optical_res ( * 3 in RGB-mode )

    * added 2 calibration-routines:
      - use last line for shading data
      - calculate average value for shading data
      - prepared routine for complex calculation

    * added gamma download curve format type 0 and 1

    * defined gamma DCF = 1 for UC1200S and UC1260

    * defined gamma DCF = 0 for UC1200SE and ARCUS PLUS

    * changed sense_handler: if bad sense_key => DEVICE_BUSY
      problem with black area at top of scan for some SCSI-adapters is solved now

    * changed wait_scanner: waits 0.5 sec now

    * disabled quality calibration in preview mode

    * made upper_left_x and upper_left_y even values
      it looks like even resolutions > 600 dpi work now!
      (not ok!!!)

    * added sigterm_handler for umax_reader_process
      changed sane_canel, do_canel and do_eof(=close_pipe),
      cancel now works!

    * changed trim_rowbufsize that made problems with SGI,
      there has been an overflow because of big buffersize


sane-0.74 -> 1.00 (build 0)
---------------------------

    * added output of umax-backend-version in sane_init

    * added support for Linotype Hell Jade2 (Vobis ScanBoostar Premium)
      SCSI-ID "LinoHell","Office2"

    * changed base of calculation of scanwidth and scanlegth from
      pixel to resolutionbase (1/300, 1/600 or 1/1200 inch)
      problem with wrong pixelcount at higher resolutions should be
      solved now

    * changed selection for calibration mode

    * changed calculation for width in pixels for lineart and halftone mode

    * added inquiry for UMAX UC1200SE

    * corrected 12/36-bit-mode in send_gamma_data and array in umax.h

    * added support for Gemini D-16


sane-1.00-build-0 -> build-1
-----------------------------

    * added support for Nikon AX-210

    * changed all grey to gray (David Mosberger asked for it)

    * changed bug in gray gamma table conversion

    * added Linotype Hell Saphir 2 as supported scanner

    * changed umax-uc1260.h

    * added "- 1" in umax_order_line:
      else if (dev->pixelline_opt_res < length * 3 - CCD_distance * 3 - 1)
      else if (dev->pixelline_opt_res < length * 3 - CCD_distance - 1)

    * changed all allloca() to malloc() + free()

    * removed options speed and smear, I got the info from umax that
      these values should never be changed!

    * removed ten_bit_mode and twelve_bit_mode options, added option bit_depth
      and added support for 9, 14 and 16 bpp

    * replaced u10_range and u12_range by dynamically set scanner.output_range

    * removed static allocation for gamma_table[0..3], this is done dynamically
      in sane_open and freed in sane_close now.

    * added an empty string to the end of scan_mode_list and source_list

build-1 -> build-2
------------------

   * added "*info |= SANE_INFO_RELOAD_PARAMS;" in control_option set bit_depth

   * added inverted calibration modes for Astra 6X0S scanners.


build-2 -> build-3 (sane-1.0.1)
-------------------------------

   * inversion of calibration is automatically set for Astra 6X0S and Vista S6E

   * check for machine endianness is done only at backend start now!

   * made umax_do request_sense static


build-3 -> build-4
------------------

   * Added option name and parameter in debug output in option set value

   * Added option name in debug output in option get value

   * Changed calculation of pixels/line and lines, may be the problem
     with shaered images at resolutions>600 dpi is solved now
     later Remark: This error does still occur, but not so often!


build-4 -> build-5
------------------

   * Changed calculation in umax_order_line, color images should
     have the correct length now for Astra 6X0S and Vista S6E

   * Added DOR resolutions and area


build-5 -> build-6
------------------

   * Bits for ADF and UTA in driver internal inquiry block (old UMAX scanners)
     are set dynamically now.

   * Added Genius ColorPage-HR5 (vendor KYE) (=Astra 1200S) as supported scanner

   * Added Astra 2400S as experimentally supported scanner

   * Added pauses for Astra 2400S

   * Removed reposition_scanner at start of scan, Astra 2400S does not work with it!

   * changed debug output in correct_inquiry

   * for UMAX S12 the driver does not wait any more until scanner is repositioned
     this is experimental, it may work with other scanners too!


build-6 -> build-7
------------------

   * changed some int -> unsigned int (resolution and scanwidth etc)

   * calculate_pixel is called with right values in UTA mode now.

   * if DOR resolution is not defined and DOR is available
     the DOR resolution is set to double standard resolution

   * removed wait_scanner after set_window-command

   * added Mirage D16-L as supported scanner

   * Astra 2400S works stable now


build-7 -> build-8
------------------

    * for Mirage D-16L the driver does not wait any more until scanner is repositioned
      this is experimental, it may work with other scanners too!

    * Mirage D-16L does not give back information for calibration by driver,
      added new modus that "guesses" the values for Mirage D-16L.

    * Added read_shading_data to 16 bit calibration routine and made a routine for
      averaging with low byte first and one for high byte first

    * Byte order (HBT) is set always (also in lineart and halftone)
      because it is used in driver calibration

    * Added relevant_*_res the is set to normal or double resolution values

    * Added test if scsi buffer is large enough for one shading line

    * selection for Calibration mode, Shading type and Speed/Smear
      are only compiled in if define is set in umax.h


build-8 -> build-9
------------------

    * slow speed options is preset for Vista T-630

    * speed/smear can be set now (forgotten #ifdef)

    * selection of exposure times now is only available if it is enabeld

    * gamma table is disabled in lineart and halftone modes

    * if image is transmitted completely sane_read (close_pipe)
      returns with SANE_STATUS_GOOD in adf mode and in any other case
      with SANE_STATUS_EOF


build-9 -> build-10
-------------------

    * removed change about sane_read returns SANE_STATUS_GOOD in ADF mode

    * transformed all gif files into jpeg format because of all the
      trouble about the gif licence. Will never use gif again!

    * added correction for inquiry-ADF-bit for Astra 2400S

    * added second test method to detect ADF error in sane_start:
      inquiry block is read and cover open, jammed and no paper is tested


build-10 -> build-11
--------------------

    * test for adf errors (sane_start) is now only done if adf is selected


build-11 -> build-12
--------------------

    * moved adf error tests from sane_start to umax_start_scan

    * sane_control_option: printing option name is set to "(none)" if pointer is NULL

    * added error message after each call of sanei_scsi_cmd

    * changed manipulation of ADF bit for Astra 2400S in umax_correct_inquiry
      from dev->adf to set_inquiry_sc_adf()


build-12 -> build-13
--------------------

    * changed inquiry data of UC1200SE

    * added set_inquiry_CBHS(50) in correct_inquiry for UC1200SE

    * correct_inquiry is called for scanners with short inquiry now

    * dev->inquiry_max_calib_lines is preset to 66,
      problem with Vist S6E V1.6 should be solved this way.

    * inquiry block parameters > 0x93 are only used if block is long enough


build-13 -> build-14
--------------------

    * changed inquiry_cbhs_[min,max] -> inquiry_[contrast, brightness, threshold,
      highlight, shadow]_[min,max] and set correct values for CBHS_255 and CBHS_50

    * changed cbhs_str[CBHS_50] from "78-178" to "0-50"

    * changed umax-uc1200se.h inquiry byte 0x64 bit 5 to 0 (gamma download format != 2)
      and wdb len from 0x30 to 0x31

    * disabled halftone pattern download, and halftone dimension selection - not ready
      preselected is a 8x8 built in halftone pattern

    * removed set_inquiry_CBHS(50) from correct_inquiry for UC1200SE

    * added "Astra 2200 " as supported scanner

    * added cast from (void *) to (unsigned char*) in umax_send_gamma_data


build-14 -> build-15
--------------------

    * add "(unsigned)" to option number test in sane_control_option

    * for Astra 1200 the driver does not wait any more until scanner is repositioned
      this is experimental, it may work with other scanners too!


build-15 -> build-16
--------------------

    * OPT_SOURCE: removed change of width/height and "*info |= SANE_INFO_RELOAD_PARAMS;"
      from get_value to set_value and added "*info |= SANE_INFO_RELOAD_OPTIONS;"


build-16 -> build-17
--------------------

    * changed calibration mode from SHADING_TYPE_AVERAGE_INVERT to SHADING_TYPE_AVERAGE
      for UMAX S-6E, UMAX S-6EG and Vista-S6E

    * changed several int -> unsigned int

    * added support for sanei_scsi_open_extended and variable scsi buffer size


build-17 -> build-18 (sane-1.0.2)
---------------------------------

    * some cosmetic changes of the form
        if (..) {...}
      =>
        if (..)
        {
          ...
        }

    * removed unnecessary SHADING_TYPE-handling routines for UMAX S-6E, UMAX S-6EG and Vista-S6E

    * scsi buffer is only reallocated when size has changed


build-18 -> build-19
--------------------

    * added UMAX PowerLook III as supported scanner

    * added EPSON Perfection600 as supported scanner (is a Astra 1200S)

    * type of option 0 (Option numbers) is set to SANE_TYPE_INT now (undefined before)

    * removed bug that mixed 12 bit and 14/16 bit gamma table when 14 or 16 bit are supported
      and changed value range of grey custom gamma table:
      - maximum output value is 255 for 8 bits/pixel
      - maximum output value is 65535 for 9-16 bits/pixel
      the custom gamma table does work now with 9-16 bits/pixel output mode

    * added routine that reverses bit order in lineart mode when scanner sends LSB first in lineart


build-19 -> build-20
--------------------

    * added variable calibration_width_offset, Mirage D-16L offset of 308 is defined there
      and added offset of 28 for PowerLook III

    * umax_do_calibration: guessed calibration data values use 2 bytes/color if scanner
      supports gamma input bit depth supports 2 bytes/color

    * calibration by driver now does work for PowerLook III

    * added function umax_set_rgb_bind() that sets the sensitivity of some options
      that make available different settings for RGB. This function now is called
      when scanmode is selected.

    * changed some int to unsigned int

    * made dev->pause_after_reposition = -1; as default, so as default sane-umax does not wait
      any more while the scanhead is moving back - this may break the support of some scanners
      so it may have to be changed (dev->pause_after_reposition = 0;) for some scanners.
      reported to work for:
      - Supervista S12
      - PowerLook III
      - Astra 2200


build-20 -> build-21 (sane-1.0.3)
---------------------------------

    * Added Linotype "JADE" as supported scanner

    * Removed bug in sane_close that caused segfault when
      sane_close was called twice with the same scanner handle

build-21 -> build-22
--------------------

    * replaced fgets() by sanei_config_read()

    * added command queueing for scsi read commands
      - removed WAIT_SCANNER scsi command call after read command
      - the additional buffers are allocated in the reader process
      - the number of buffers can be defined in umax.conf
        with option scsi-maxqueue [1..8] default:2
      (did some tests with Supervista S12 and PowerLook III
       on the same scsi controller (ncr53c810):
       command queueing works, it is possible to start
       both scanners at the same time. The scanner with the
       higher scsi-ID (=priority) does scan without stops,
       the other scanner has to wait long time between the
       scsi-read commands, it also is possible to give
       both scanners different scsi command queueing depth).

    * prepared replacement of pipe by sanei_ipc_* routines

    * added routines umax_get_lamp_status and umax_set_lamp_status

    * updated inquiry output (new bits from powerlook scsi protocol 1.0 X101)

    * added new (powerlook scsi protocol 1.0 X101) window descriptor block entries to umax-scsidef.h

    * added button protocol (3 buttons, from powerlook scsi protocol 1.0 X101)

    * added button "lamp off" to turn off scanner lamp
      (this option may be remoeved again in later version, may be
       it will be replaced by an autmatic lamp turn off function)

    * added options to umax.conf
      lamp-off-at-exit [01] default:0
      lamp-off-time [10..720] (in minutes) default:30

    * if  lamp_off_at_exit == 1 and lamp_control_available == 1
      the scanner lamp is turned off in sane_close

    * renamed umax-scanner.h and umax-"scannername".h to umax*.c
      because these files create data code


build-22 -> build-23
--------------------

    * scsi commands set/get lamp control are send with a command length of
      10 bytes instead of 6 bytes and control bit #7 is in byte 0x03
      and not in byte 0x04 (it is described incorrect in the umax doc)

    * enabled lamp control for Astra 2200

    * removed umax.conf options lamp-off-at-exit and lamp-off-time

    * added frontend option lamp-off-at-exit

    * added frontend option lamp-on

    * forcing calibration_bytespp = 1 for Astra 2200 - quality calibration does work now

    * for Astra 2200 selection for different x/y resolutions is disabled


build-23 -> build-24 (sane-1.0.4)
--------------------

    * forcing calibration_bytespp = 1 for Astra 2100S

    * Added Astra 2100S as supported scanner

    * enabled lamp control for Astra 2100S

    * Added warning in sane_control_option if option could not be set


build-24 -> build-25
--------------------

   * dev->lamp_control_available is preset to 0

   * dev is erased (0) after it is allocated

   * added special options for PowerLook 3000:
     - calibration is done for each CCD pixel
     - calibration_bytespp is set to 1
     the scanner still is not listed as supported scanner because
     the changes are not tested

   * added options that are read from umax.conf:

     - execute-request-sense:
       values: 0=disabled, 1=enabled, default=0
       If set to 1 umax_do_request_sense is called in umax_do_calibration.
       This can hang the system (but has been enabled until this version)

     - scsi-buffer-size-min, scsi-buffer-size-max:
       values: 4096-1048576, default min=65536, max=131072

     - scan-lines, preview-lines:
       values: 1-65535, default scan-lines=40, preview-lines=10
       define the maximum number of lines that are scanned into one buffer

     - force-preview-bit-rgb:
       values: 0=disabled, 1=enabled, default=0
       set preview bit in rgb real scan

     - slow-speed, care-about-smearing:
       values: -1=auto, 0=disabled, 1=enabled, default=-1
       dangerous options, needed for some scanners

     - calibration-full-ccd:
       values: -1=auto, 0=disabled, 1=enabled, default=-1
       do calibration for each pixel of ccd instead of selected image

     - calibration-width-offset:
       values: -1=disabled, >=0 set value
       add an offset width to the calculated with for image/ccd

     - calibration-bytes-pixel:
       values: -1=disabled, 0=not set, 1=1 byte/pixel, 2=x bytes/pixel
       use # bytes per pixel for calibration

     - invert-shading-data:
       values: -1=auto, 0=disabled, 1=enabled, default=-1
       invert shading data before sending it back to the scanner

     - lamp-control-available:
       values: 0=automatically set by driver - if known, 1=available, default=0

   * updated html-documentation

   * updated sane-umax.man

   * added pre focus options (PowerLook 3000):
     - FP  - fix focus position
     - FDC - lens calibration in doc position
     - PF  - disable pre focus
     - LCL - use 0.6/0.0mm holder focus position

   * added #define SANE_I18N(text) text to umax.c. This is to mark the backend
     specific texts that shall be translated by the frontend


build-25 -> build-26
--------------------

   * added umax.conf option handle-bad-sense-error #

   * added Mirage II as supported scanner

   * added routine umax_set_max_geometry that defines available surface
     and resolution in dependence of scansource and double optical resolution
     (before this bugfix the change of dor mode overwrote the settings of
      scansource and vice versa)

   * added ESCORT Galleria 600S as supported scanner, is an Astra 600S

   * removed if (config_line[len - 1] == '\n')... while reading config file

   * changed some static unsigned int to static int

   * sane_exit does free devlist

   * added some options for Powerlook 3000 - still is not completely supported


build-26 -> build-27
--------------------

   * added Linotype Saphir4 as supported scanner (is a Powerlook III)

   * increased length of vendor, product and version by 1 to make sure the
     string always ends with '\0'

   * changed default minimum scsi buffer size from 64KB to 32 KB


build-27 -> build-28
--------------------

   * removed "#include <assert.h>" from umax.c, assert() is never called

   * added UMAX Powerlook 3000 as supported scanner

   * bugfix for 3 pass scanning:
     reposition_scanner is called in sane_start for 2nd and 3rd pass of 3 pass scan


   * bugfix for UTA and DOR geometry offsets:
     - added inquiry_uta_x_off and inquiry_uta_y_off
     - geometry ranges are defined from xoff..xoff+dx, yoff..yoff+dx now
     - check_values also does test for this range
     - removed dev->maxwidth and dev->maxlength, values are calculated in place now

   * bugfix for DOR:
     - bottom-right edge keeps in position if possible when DOR is switched on

   * bugfix for Astra 600S and 610S color mode:
     - scale_y is set to 0.5 if y_resolutions is smaller than optical_res/2
     this does fix the color scan problems for Astra 610S.

   * added Heidelberg Linoscan 1400 "HDM LS4H1S" as supported scanner


build-28 -> build-29
--------------------

   * removed bug in umax_set_max_geomentry: now left y keeps in position when possible

   * rgb bind keeps unchanged when scanmode != color (only gray selection active in this case)

   * in function static int umax_give_scanner(Umax_Device *dev) moved umax_reposition_scanner(dev);
     to the end of the function

   * set calibration area for astra 2200 to full ccd and calibration_bytespp to 2

   * UC840 does wait in reposition_scanner until scan head stops moving

   * set minimum resolution to 5 (was 1 before what could make problems)


build-29 -> build-30
--------------------

  * added EDGE KTX-9600US as supported scanner

  * Changing scansource does not change lower left selection edge
    if not necessary


build-30 -> build-31 (sane-1.0.6)
---------------------------------

  * sane-umax.man uses @LIBDIR@ and @CONFIGFIR@

  * changed automatic calibration_width_offset marker from -1 to -99999

  * changed
      dev->calibration_area to 1 and
      calibration_width_offset to -1 and
      inquiry_max_calibration_data_lines to 66
   for UMAX Supervista S-12, LinoHell Office and LinHell JADE


build-31 -> build-32 (sane-1.0.7)
---------------------------------

  * added Nikon AX-110 as supported scanner (Vista S6E)

  * added umax.conf option "gamma-lsb-padded" with options:
    -1 = auto
    0 = inactive (msb padded)
    1 = active (lsb padded)
    activated for astra 1220s, internal 12 bit gamma correction does work now


build-32 -> build-33
--------------------

  * added USB support for Astra 2200
    thanks to Frank Zago <fzago@greshamstorage.com>
    for creating umax-usb.c and suggesting several changes to umax.c

  * added umax.conf option "connection-type", values: 1=scsi, 2=usb

  * added umax.conf device selection "usb 0x1606 0x0230" (usb vendor_id product_id)

  * removed #ifdef HAVE_SANEI_IPC parts (SANEI_IPC was never defined)


build-33 -> build-34 (sane-backends-1.0.8)
--------------------

  * removed bug in reader_process / umax_queue_read_image_data:
    - the integer value that keeps the length information is kept valid
      as long a queued scsi command is still running now
    - when the scsi command returns with less data than requested
      then the routine does handle this correct now
      (before the routine handled this as if the requested amount of
       data was returned)

  * backend waits until scanhead has repositioned after scan for UMAX UC1260

  * y_dpi quanitsation is set to 5 (same as x_dpi quantisation - was 1)


build-34 -> build-35
--------------------

  * changed ID string "LS4H1S  " to "LS4H1S " (remove one space)

  * activated lamp control for Astra 1220S V1.5


build-35 -> build-36
--------------------

  * added PowerLook 2100XL as supported scanner

  * settings for PowerLook 2100XL:
    calibration_width_offest = 52
    (calibration_bytespp = 2, default)
    (calibration area = image, default)

  * force_quality_calibration defined for PowerLook 2100XL:
    (normal calibration mode expects different calibration_width_offsets
     for PowerLook 2100XL, so it is not possible to calculate the correct
     number of calibration pixels in normal calibration mode)
    - disable quality calibration option for frontend
    - always set quality calibration bit when scanning

  * changes for exposure times:
    - Inquiry does print correct maximum value now
    - changed variable type from SANE_FIXED to SANE_INT
    - exposure times setting for colormode does work correct now
    - as default there is only one common setting for exposure times in rgb mode,
      added umax.conf option exposure-time-rgb-bind to enable different settings
      for each color
    - added option select-calibration-exposure-time, disabled by default.
      if disabled then the calibration expousre times are the same as the scan exposure times

  * added UMAX PS-2400X as supported device

  * added batch scanning support:
    Options:
    --batch-scan-start:		calibration, scan, scanhead is not moved into home position
    --batch-scan-loop:		no calibration, scan, scanhead is not moved into home position
    --batch-scan-end:		no calibration, scan, scanhead is moved into home position
    --batch-next-tl-y:		defines the y start position of the bext batch scan

    - calibration for batch scan needs a different offset and always is done for full ccd.
      PowerLook 2100XL: calibration_width_offset_batch = 1052
      Powerlook III:	calibration_width_offset_batch = 828


build-36 -> build-37
--------------------

  * added Linotype Hell SAPHIR3 as supported scanner

  * if batch-next-tl-y is not set then the backend uses br_y


build-37 -> build-38 (sane-backends-1.0.9)
--------------------

  * added Power Look 2000 as supported scanner


build-38 -> build-39 (sane-backends-1.0.10)
--------------------

  * Power Look 2000: corrected wrong calibration-width-offset from 52 to 22
    and added calibration-width-offset-batch to 24

  * initialized some global pointers and num_devices in sane_init

  * replaced fork() by thread for OS2

build-39 -> build-40 (sane-backends-1.0.12)
--------------------

  * UC630: dev->pause_after_reposition = 0


build-40 -> build-41 (sane-backends-1.0.13)
--------------------

  * Astra 2200 SU via USB connection: quality calibration is reset and not selectable


build-41 -> build-42
--------------------

  * replaced fork() routines by sanei_thread routines

  * added call sanei_scsi_req_flush_all(); to do_cancel - in threaded case this is not done when reader_process is canceled

  * in reader_process: marked freed scsi buffers as freed(NULL). This is necessary when reader_process is run as thread!

  * Backend is compatible to Mac OS-X now.


build-42 -> build-43
--------------------

  * DOR mode does test for non negative upper_left_x values, set to 0 if negative


build-43 -> build-44
--------------------

  * the sense_handler stores the sense data in buffer[0] when valid sense data is available

  * do_calibration:
	- corrected bug in reading calibration width from scanner (added /bytespp)
	- when additional request sense is not called then the buffer is not erased,
	  the sense data from the sense_handler is used
	- when additional request sense is called then the buffer is erased at first

build-44 -> build-45
--------------------

   * added default options for Linotype Opal2
   * changed bug in highlight-blue function (using green value)