* The -channel option is operational again.
* The black channel offset was incorrectly computed when compositing causing
a seg-fault (bug report from Arfune).
* Under extremely rare conditions the GIF coder had a memory leak.
* Moved image list advancement to after the progress monitor call in
QuantizeImages() (bug report by Tim Hunter).
* The black channel offset was incorrectly computed when compositing causing
a seg-fault (bug report from Arfune).
* Under extremely rare conditions the GIF coder had a memory leak.
* Moved image list advancement to after the progress monitor call in
QuantizeImages() (bug report by Tim Hunter).
* The display program no longer exits after the first image.
* The JP2 coders no longer interprets the first channel as an opacity channel.
* The VID coder no longer generates a fault (bug report from Arfune).
* CMYK Postscript with a subimage specification no longer returns a grayscale
image (bug report from Arfune).
* Handle image settings with MogrifyImageInfo() method.
* GNS contributed QueryMultilineTypeMetrics(). Its like QueryTypeMetrics()
but returns the maximum text width and height for multiple lines of text.
* Treat -channel as a setting in the command line programs.
* The black channel offset was incorrectly computed when compositing causing
a seg-fault (bug report from Arfune).
* Under extremely rare conditions the GIF coder had a memory leak.
* Moved image list advancement to after the progress monitor call in
QuantizeImages() (bug report by Tim Hunter).
in the process. (More information on tech-pkg.)
Bump PKGREVISION and BUILDLINK_DEPENDS of all packages using libtool and
installing .la files.
Bump PKGREVISION (only) of all packages depending directly on the above
via a buildlink3 include.
All library names listed by *.la files no longer need to be listed
in the PLIST, e.g., instead of:
lib/libfoo.a
lib/libfoo.la
lib/libfoo.so
lib/libfoo.so.0
lib/libfoo.so.0.1
one simply needs:
lib/libfoo.la
and bsd.pkg.mk will automatically ensure that the additional library
names are listed in the installed package +CONTENTS file.
Also make LIBTOOLIZE_PLIST default to "yes".
* PNG vulnerability fix.
* Added YCbCr coder for raw Y, Cb, and Cr samples.
* Fixed the -fx option operator precendence problem (+ and - should be the
same precendence).
* Rare memory corruption due to a JPEG comment (bug report by
brundlefly76@hotma...).
* The PerlMagick Evaluate() method no longer fails with an assertion error.
* Restore -dSAFER for a more secure Ghostscript invocation. The downside is
we must copy the Postscript file to a temporary file otherwise we get a
the ocassional invalidfileaccess fom Ghostscript.
* Do not report image depth unless the -verbose or -format %z option is used
with the identify program.
* Center gravity now works when cropping by percent (e.g. -crop 75%).
* Small memory leak on exception in the ImageToBlob() method.
* %@ returns the image bounding box (e.g. identify -format %@ logo:).
* Sync pixel cache when cloning an image-- just in case someone clones
an image clone and accesses pixels with AcquireImagePixel() instead of
GetImagePixels().
Changes:
Bug fixes:
* Fixed problem with relative filename wildcards (e.g.
identify images/*.jpg)
* DrawSetStrokeDashArray() now respects the number_elements argument.
Performance improvements:
* Don't open a loadable module if its already open.
* Added anonymous memory-mapping to the pixel cache. This makes it
possible for memory to be returned to the system after the image is
destroyed.
Utilities enhancements:
* The display program now composites any transparent images on top of a
checkered background to assist in visualizing any partial transparent
pixels.
Coder additions/improvments:
* When an unknown property is encountered in the XCF coder, it enters a
bogus loop that never terminates.
Code/structure enhancements:
* Introduce MagickBooleanType and MagickStatusType types of unsigned int.
These types are 100% compatible with the existing API but removes
ambiguity. Previously we got complaints that users did not realize that a
return value of unsigned int in most cases meant a binary value (0/1).
MagickStatusType is used for methods that return a bit mask.
* Introduce MagickFalse and MagickTrue while we deprecate True/False.
And more...
* DrawSetStrokeDashArray() now respects the number_elements argument
* Since the WMF coder has Wand dependancies we only enable this coder when
both WMF is available and module support enabled
* Some XCF tiles were not being read properly
* Fixed memory leak in EPT coder
* Fixed memory leak in DestroyImage()
* Close Windows registry when no longer needed
* RelinquishUniqueFileResource() was not freeing file resource properly
* Free the ImageInfo structure allocated by CloneImageInfo() in the
WriteImages() method
* EPT images cannot support blobs
* Fixed temporary file resource limit leak in RelinquishUniqueFileResource()
* ConvolveImageChannel() incorrectly initialized non-selected channels to 0
* Added Cisco IP phone image format (write only)
(to ignore any errors).
This is so in the rare case that you already have files from an
old install that it won't fail to install new package.
This was discussed with maintainer, tron@, in March.
* MagickWand API, a higher level C interface to the ImageMagick methods.
* Increased stability. The library was checked and cross-checked with
splint and valgrind to eliminate security vulnerabilities and common
programming mistakes as well as ensure no memory leaks.
* A new -fx option to apply a mathematical expression to an image list from
the command line.
* Algorithms were improved to reduce processing requirements resulting in a
significant speed-up over previous releases.
* All regression tests pass at Q8, Q16, and Q32 on a variety of platforms
including Linux, Solaris, FreeBSD, Mac OS X, and Windows.
* The composite methods have been corrected to precisely conform to the
SVG standard.
* New image comparison utility, `compare'.
* Better support for CMYK image workflow.
* New string handling methods to help prevent potential buffer overflow
problems.
* Error reporting has more detail to help identify the cause of a
particular problem.
* Bug fixes
by moving the inclusion of buildlink3.mk files outside of the protected
region. This bug would be seen by users that have set PREFER_PKGSRC
or PREFER_NATIVE to non-default values.
BUILDLINK_PACKAGES should be ordered so that for any package in the
list, that package doesn't depend on any packages to the left of it
in the list. This ordering property is used to check for builtin
packages in the correct order. The problem was that including a
buildlink3.mk file for <pkg> correctly ensured that <pkg> was removed
from BUILDLINK_PACKAGES and appended to the end. However, since the
inclusion of any other buildlink3.mk files within that buildlink3.mk
was in a region that was protected against multiple inclusion, those
dependencies weren't also moved to the end of BUILDLINK_PACKAGES.
