19991002 ajv created CVS repository
19991002 ajv Started work on debugging issues
19991003 ajv Moved main() to getopt for maintainability
19991003 ajv Added all the support files, and ported the
19991003 ajv Removed multiple target support, debugged other printers
Made recompilation to change things unnecessary
19991013 ajv Added internationalisation support
Added English man page
Reduced getopt() to *BSD friendly version
Fixed a bug that ajv found, and Giorgio fixed
Updated the usage to include build version and all
19991020 ajv Added back '-' and post-option file parsing
Added I18N stuff
Added "verbose" option to output to stderr
19991021 giorgio Added all of Giorgio's patches
Added X & Y offsets
19991021 ajv FreeBSD fixes in Makefile
Documentation update, including man page
19991105 ajv Documentation update
Dutch support - Jimmy Tak
"letter" added as valid paper size - Ionutz Borcoman
RedHat doco updated - Duncan Haldane
19991120 giorgio Much faster output
Much better printing in dark regions
compression reenabled for black mode
19991127 giorgio Changed back to standard Cartesian distance for colors
Gamma correction tool (see separate archive)
bug fixes in black printing
19991128 ajv re-merged with Giorgio's code base
All patches from 19991105 are back
updated shift sizes in pnm2ppa.conf
fixed permission problem with make install
fixed "verbose" and "dump" pnm2ppa.conf keywords
19991204 duncan "legal" (US Legal, 8.5in x 14in) added as a valid paper
size - Duncan Haldane <firstname.lastname@example.org>
20000113 ajv broke out image.c into several files in an attempt
to add pgm support without major breakage
20000114 ajv Finally fixed annoying memory bug. Both MAXLINES and
Max Buf Lines must agree.
20000120 klamer Patches to allow calibration to work again.
mancini BeOS support
duncan RedHat Print filters added to CVS tree
ajv Fixed -v and moved yet another global into a struct
20000306 duncan Lots of little fixes; --fd working right now;
bidirectional shearing adjustments added (user
configurable entries "blackshear" and colorshear"
20000311 duncan added : -B n option and "blackness n" config file
entry to control black ink density (drops per
pixel n = 0,1,2,3,4 ) default is 2. Also there are
now config file entries blackshear and colorshear
for adjusting shearing in bidirectional printing.
Fixed various bugs that broke HP820/1000 printing.
20000318 duncan Color EcoMode printing now uses bidirectional
printsweeps (formerly only FastMode did this)
If gMaxPass = 1, you now get bidirectional sweeps
(in cut_im_color_swath() , cutswath.c)
20000326 duncan ** changes to dither.c:
on odd lines, black is
now processed (not color), even black pixels are
duplicated, so if either of two successive pixels are
black, they both are. Changes also fix Tim's report
about blurry black/white boundaries.
** changes to image.c
If the odd line is not printed
(PixMap or EcoMode), image.c now stores the line, and
adds the black parts to the next even line.
This fixes a bug in where thin (one-pixel wide) black
lines (like enscript borders) could be lost because
black ink only printed every second pixel, and odd
black lines were just copies of the preceeding even
line which had been stored.
20000331 duncan added support for large papersizes 11x17 and a3
(for the 1000 series printers). Added a
printer-specific max_pagewidth so papersizes that
are too wide for a printer series will not print
on that printer.
20000402 duncan tweaked bidirectional sweep shearing correction.
(removed the +1 right-to-left black shift in
cut_im_black_swath() in cutswath.c ).
I am not sure if the shearing correction is
model or individual printer-specific. (I suspect
the second case, in which case it is a calibration
issue (The HP Win9x software seems to have calibration
adjustments for blackshear.) if any corrections
turn out to be generic to a given printer family,
(rather than specific to a particular printer)
the default can be changed in defaults.h (the
20000404 duncan added support for pbmraw and pgmraw formats.
all three pnmraw formats now work.
(changes in image.c, image.h )
20000405 duncan added support for pgm, ppm, pbm nonraw formats.
added papersize autoconfiguring from the pnm
header, so all ghostscript papersizes are now
supported (if the printer can accomodate them)
The -s option now only controls the gammafile
creation papersize. Made bidirectional
printing the default (but this can be configured
with a "unimode 1" in pnm2ppa.conf, or --uni,
--bi command line options.
20000408 duncan graceful termination of print run if pnm
input is truncated (i.e., if read_line()
encounters EOF). printer no longer hangs.
Moved text strings to lang.h/text-en.h
20000423 duncan clean up way things near sides of swath are
handled in cutswath.c
20000710 duncan in FS_Color_Dither(), use line= (1,1,1)
not (0,0,0) for black replacement of dark
pixels for which no color match is found.
(fixes flashing LED on certain images - those with
no black - with --eco on HP820)
(0,0,0) = black ink, (1,1,1) = gMaxPass drops of
20000928 duncan in cutswath.c, if the vertical position
of the black and color sweeps differed by 1 , 2 or 3
dots, HP820 got "flashing light syndrome".
Fixed by testing for this possibility, and
if found, by reducing numlines, and padding the top
of a swath with an (even) number of blank lines
to keep the height the same.
(Thanks to Steven Elliott <email@example.com>
for reporting this bug, and testing the fix.
20001010 ajv made input of paths to gamma and config files secure
against buffer overruns.
duncan In -g mode, check that gamma.ppm is not a symbolic
link, and is a regular file, before writing to it
(but after opening it).
20001018 duncan clean up calibration program. now called
calibrate_ppa. It does most calibration now.
moved the job of writing gamma.ppm for color
calibration to a new utility calibrate_ppa.c.
pnm2ppa -g now just prints the corresponding
secured syslog messages. No user input is
sent to syslog. syslog messages are wrapped, and can
be suppressed with a new keyword "silent 1"
in /etc/pnm2ppa.conf (but not in subsequently-read
config files). all syslog activity now occurs
20001106 reduce memory footprint of calibrate_ppa
(it was 36MB!).
20001112 fix syslog silent/verbose usage; fix alignment
pattern bug in calibrate_ppa. repaired the HP820
fix that got broken when the vertical offset was
redefined by 600 dots...
20010829 duncan Change the preprocessor flag __BEOS__ to __NO_SYSLOG__
as it is not BeOS specific (it supresses use of the
System Log) Added Makefile entries for Solaris, and
an INSTALL.SOLARIS file.
20020211 duncan Cleanup C++-style comments (gives gcc compiler warnings)
Small syslog.c/Makefile cleanups for Solaris.
Minor documentation updates. Change a confusing
syslog message about "unsupported PNM formats" so it is
clear that the error was non-PNM input to pnm2ppa.
Increase version to pnm2ppa-1.10
20020224 duncan Reworked cutswath.c and ppa.c so print sweeps
are sent separately by ppa_print_page()
in cutswath.c to ppa_print_sweep() in ppa.c
(until now, they were going in groups of
gMaxPass= 4 at a time in color high quality mode)
This will now permit improvement of the print sweep
pattern, and make the code easier to understand
(easier to separate ppa stuff from image stuff)
if someone wants to do a gimp-print PPA backend.
20020311 duncan lots more (incremental) reworking of cutswath.c;
now it's pretty well modularized, reading the input
image constructing the sweep data, constructing
the nozzle data, are all done in separate functions
now. Also, modified image.c so the
the expansion of data to ppm format
(three bytes per 8 dots) is not done for
pgm or pbm input (unless the -p option which uses
colored ink to make black, and disbles the black
ink is used). a new global variable gByteCount = 1
or 3 distingushes the two types of intermediate data
Also still TODO: I finally understood that
the striping of black parts of the image when
e.g. the gs example file "tiger.ps"is printed in
high-quality mode is because the color and black
sweeps are not aligned, so sometimes the color prints
on top of black, sometimes the color prints
underneath black. If both types of ink are to be used
at the same place, the simple solution is probably to
prevent black sweeps getting ahead of
unprinted color in ppaprint_page() in cutswath.c.
20020526 duncan solved the striping problem; reworked cutswath.c
so black ink is never overprinted by color ink.
20020601 duncan reworked the code that prevents color swaths
and black swaths from having vertical_pos that
differs by 1, 2 or 3 lines (needed for HP820C).
(to go with the new swath-cutting algorithm).
Added a method for recognising if input is a
calibration page produced by calibrate_ppa,
as these dont work correctly with the new
swath-cutting algorithm. (will restore
a "legacy-style" (pnm2ppa-1.04-style) method
for printing input that is detected to be
20020610 duncan added a method for storing info on lines as
they are read from the image. (whether they have black,
and/or color data, etc.) Helps a lot!
final fixes to calibrate_ppa. new code seems to work
20020623 duncan added a new command line option --dpi300 that
interprets input resolution as 300dpi instead of 600dpi
20020930 duncan enable compression of (color) data by default (only
compression of black data was previously enabled).
added a configuration file keyword "compression"
to switch it off, in case there are problems ...
(color data compression seems to work on HP722C!)
20021005 duncan disable logging of informational syslog messages by
default. They can (only) be reenabled by "log_info 1"
in the default configuration file /etc/pnm2ppa.conf.
(They will still go to stderr in "verbose" mode).
20021109 duncan make "version 0" in the config file set the printer
type to NOPRINTER, and abort printing if this is not
subsequently set to a valid printer model.