2/24/93 (JPEG Version 1.1)

BUG FIXES:

Changed the fill marker handling.  On early JPEG revisions, one
interpretation is that fill markers came in pairs of 0xff 0xff, but
now it can be an arbitrary number of 0xff bytes.  A one line fix in
stream.c handles this.

The arbitrary frame width and frame height is handled much more
robustly.  Decimated frames default frame size is handled with
rounding up rather than rounding down to keep compatiblity with the
JPEG CD.  This has changes in marker.c and jpeg.c and io.c.

CODE:

(marker.c) (io.c) (jpeg.c) Setting the Framewidth and Frameheight that
violate logical MDU blocks will flag warnings.  The frame rounding has
been changed from rounding down to rounding up. Rounding up preserves
MDU properties and is consistent with the CD specs.  (We do not round
down anymore).  The previous Framewidth and Frameheight does not work
with values that cross MDU boundaries; It now works arbitrarily.

  1) In decoding, if any space is created, the space is set to "zero" bytes
  (depending on the implementation of disk file "holes").

  2) In encoding, if any space is required, the space is set to pel
  replication of the boundaries.

(jpeg.c) The default filename is now componentfile0.jpg rather than
componentfile0.J.

DOCUMENTATION:

The colorspace section was edited to include more precise matrices and
a description of CCIR-601 versus JFIF style YUV->RGB conversion.

Some printers (non-laserwriters) have been unable to cope with the
doc.ps files, specifically problems occurred in the postscript include
files.  For some reason, Apple laserwriters and ghostscript do not
seem to be afflicted by the above problem.  We have found a change in
the dvi2ps postscript header that seems to fix this problem for Sun
newsprint (and probably other) printers as well.  If this still is a
problem, the raw tex and .ps files are available in the file
JPEGDOCv1.1.tar.Z.  Caution: the special scaling command has different
relative scales for different dvi2ps, dvips programs - can be a factor
of 100 off depending on the version of the header file.

6/14/93

Changed a few things in the documentation to reflect the YCbCr
terminology.  Other than that, essentially the same.  Added some
declaration for SYSTEM V libraries and libraries without ftruncate().

Version(1.2)

7/28/93

Changed the lossless predictor encoding functions so that 16 bit
accuracy is attained.  Some problems occurred with improper modulo
arithmetic when full 16 bits were used with certain predictors.  The
changes were in codec.c Lossless*() and a couple of lines in jpeg.c
which used &0xffff to properly trim the values.