Some part of the revbump script obviously did the wrong thing here and
replaced ${RUBY_PKGPREFIX} with whatever it expands to on the developer's
local machine.
[It is worth building packages with PKG_DEVELOPER=yes set before committing
updates.]
GraphicsMagick 1.3.31 added PKG_CONFIG, PKG_CONFIG_PATH, and presumably
PKG_CONFIG_LIBDIR to its build system, which then appear in the output
of "gm version". Ignore the WRKDIR which appears listed there.
Changelog:
Overview of Changes from GIMP 2.10.6 to GIMP 2.10.8
===================================================
Core:
- Use adaptive chunk size in GimpProjection when rendering the
projection asynchronously, rather than using a fixed chunk size.
This provides a better trade-off between throughput and
responsiveness dynamically, based on how fast the processing is.
Tools:
- In all selection tools, show error on attempt to
subtract-from/intersect-with empty selection.
- Fix text along path not working with vertical text.
- Fix Text tool's frame position when undoing a move operation.
- Streamline Text tool's drawing blocking/unblocking logic.
- When moving a text layer using the text tool (through alt+drag),
don't change the layer's box mode to "fixed", which is unnecessary,
since the layer's size isn't affected.
- Transform and deformation operations now maintain color for fully
transparent pixels, making unerase and curves manipulation of alpha
channel more reliable.
- All transform tools now apply changes when you save or export/overwrite
an image without pressing Enter first to confirm changes.
- Heal, Dodge/Burn, Smudge, and Convolve tools now adjust the processed
buffer and mask_buffer regions according to the changes made to the
application region, as calculated by intersecting it with the
drawable and mask extents. This fixes wrong application position
when painting on a drawable whose origin is above/to the left of the
image's origin, and there's a selection active.
Plug-ins:
- Port all plug-ins to the new iterator API in GEGL.
- Improve automatic detection of HEIC/HEIF files.
- Improve RawTherapee discovery by looking up registry key (should
become useful with RawTherapee 5.5 and more).
Usability:
- Compatibility information in the Save dialog is now more understandable.
The minimum GIMP version for the XCF file is always written down when it
is GIMP 2.8 or over. The list of features warranting the minimum version
is now listed in an expander container rather than as tooltip, which
makes it more discoverable. The warning on compression is now displayed
as its own text under the checkbox and not as additional text to the
minimum GIMP version label.
CLI:
- New self-explanatory --enable-win32-debug-console CLI option
Debugging:
- New GimpBacktrace API provides an interface for creating and
traversing multi-threaded backtraces, as well as querying symbol
information. Backends are available for Linux and Windows.
- Performance log recording now available in the Dashboard dock.
The log contains a series of samples of the dashboard variables,
as well as the full program backtrace, when available. As such,
it essentially acts as a built-in profiler, which allows us to
correlate program execution with the information available
through the Dashboard.
- New performance-log-expand.py tool decodes a delta-encoded
performance log by expanding the deltas, producing a log where
each sample (and other relevant elements) contain complete
information. The structure of expanded logs is identical to that
of delta-encoded logs, the expanded log simply has no deltas.
- New performance-log-resolve.py tool resolves symbol information
in backtraces. The logs produced by GIMP only specify the program
counter at each stack frame, providing an address-map to map
program-counter addresses to actual symbols separately. This tool
looks up each program-counter address in the address map,
incorporating the relevant symbol information directly into the
backtrace.
- New performance-log-deduce.py tool that statistically deduces the
correct thread states based on backtrace address frequency, fixing
local inaccuracies.
- New performance-log-viewer.py tool that is a viewer for GIMP
performance logs, with a sample-selection area at the top and an
information area at the bottom. The sample-selection area visualizes
the sampled variables and markers using a simultaneous set of plots,
and displays the currently selected samples. The information area
shows global information stored in the log, as well as information
specific to the currently selected samples, including variable listing
and statistics, full backtrace, and profile/call-graph information.
Translations:
- Updated translations: Danish, Dutch, Finnish, German, Hungarian,
Italian, Polish, Portuguese (Brazil), Spanish, Swedish, Ukrainian.
Changelog:
GEGL-0.4.12 2018-10-23
GeglBuffer
~~~~~~~~~~
Fix fix of scaled blitting, the fix from last release introduced scaling
artifacts for all formats that are non-8bit.
Contributions to this relase
~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Ell, Øyvind Kolås and Rodrigo Lledó
GEGL-0.4.10 2018-10-16
----------------------
GeglBuffer
~~~~~~~~~~
Fix integer overflow issues for u32 blit scaling code paths.
Reduced memory clearing overhead for temporary buffers during blit.
Round alpha component during u8 bilinear/box filtering (this caused artifacts
on OSX through interaction without unknown platform differences in
libc/compilter/graphics stack)
gegl_buffer_dup now matches source buffer abyss, shift and tile-size more
closely.
Added GEGL_TILE_COPY command for backends, for more efficient buffer
duplication/copies.
New GeglBufferIterator API.
Move type-generic scaling algorithms to GeglBuffer folder, move opencl
integration bits out of buffer.
Operations
~~~~~~~~~~
gif-load: new operation for decoding GIF images/animations, with frame delay
exposed as an out-property.
exr-save: do not hard-code a version name space
long-shadow: add 'Fading (fixed length)' style with a new user-defined
'Midpoint' parameter
maniuk06: use double precision, permitting to work on larger images.
watershed-transform: improved description
unsharp-mask: permit std-dev radius of up to 1500px (from 300).
Updated all operations to use new iterator API
Contributors to this release
~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Alexandre Prokoudine, Anders Jonsson, Debarshi Ray, Ell, Hussam Al-Tayeb, Marco
Ciampa, Massimo Valentini, Øyvind Kolås, Piotr Drąg, Rodrigo Lledó and Tim
Sabsch.
Changelog:
2018-10-05 babl-0.1.60 </dt><dd>
Improved thread safety, acceleration for R'G'B'A u8 -> cairo-ARGB32 conversion.
</dd><dt>
2018-10-05 babl-0.1.58 </dt><dd>
Preserve color of transparent pixels in conversion to premultiplied alpha,
Added single precision code-paths for faster fallback/reference conversions.
New BABL_ICC_INTENT_PERFORMANCE bitflag for combining with intent as bitflags,
use of matrix+trc when relative colorimetric CLUT is present. New color model
and formats, CIE xyY.
CHANGES IN 5.2.5
NEW "set pm3d depthorder base" sorts pm3d quadrangles by projecting to z=0
NEW "set jitter vertical" displaces y coordinate rather than x coordinate
NEW array size can be determined automatically from the initializer
CHANGE place titles along x axis in plots with columnstacked histograms
CHANGE equivalent slope constraint for mcs splines at both ends of the range
CHANGE treat imaginary values plotted from a using spec as UNDEFINED (NaN)
CHANGE allow "reset" between plots in a multiplot layout
CHANGE Deprecate linux and vgagl terminals (to be removed in 5.3)
CHANGE placement of axis and tic labels in 3D projections on to xz or yz plane
CHANGE default to ./configure --without-wx-multithreading
FIX parametric function plots did not work with logscale x (regression in 5.2.0-4)
FIX polar mode "set trange" was assumed to use radians, now it tracks "set angle"
FIX clip polar grid lines and ticks to x/y range limits
FIX clipping of plot "with lines" when axes are nonlinear (regression from 5.0)
FIX clipping of all elements in finanacebars/candlesticks/boxplots
FIX clipping of 3D splot "with labels"
FIX strange interaction of "noautoscale" with blank data lines
FIX alignment of boxed text to center for eps/cairolatex
FIX incompatibility of "pm3d depthorder" and rgb color taken from data column
FIX aqua terminal font changes in enhanced text mode
version 1.0.1
This is a binary compatible release.
* lossless encoder speedups
* big-endian fix for alpha decoding
* gif2webp fix for loop count=65535 transcode
* further security related hardening in libwebp & libwebpmux
* miscellaneous bug & build fixes
Bump PKGREVISION to nb3. Despite only 35 minutes passed between the
PKGREVISION++ and PKGREVISION-- at least a bulk builder catched it leading to
failures of building x11/gtk3. Bump it to nb3 in order to be on the safest
side.
Thanks to <jperkin> and <joerg>! (Possible further regressions are mine though!)
1.3.31:
Special Issues:
Firmware and operating system updates to address the Spectre vulnerability (and possibly to some extent the Meltdown vulnerability) have substantially penalized GraphicsMagick's OpenMP performance. Performance is reduced even with GCC 7 and 8's improved optimizers. There does not appear to be anything we can do about this.
Security Fixes:
GraphicsMagick is now participating in Google's oss-fuzz project due to the contributions and assistance of Alex Gaynor. Since February 4 2018, 292 issues have been opened by oss-fuzz and 279 of those issues have been resolved. The issues list is available at https://bugs.chromium.org/p/oss-fuzz/issues/list under search term "graphicsmagick". Issues are available for anyone to view and duplicate if they have been in "Verified" status for 30 days, or if they have been in "New" status for 90 days. There are too many fixes to list here. Please consult the GraphicsMagick ChangeLog file, Mercurial repository commit log, and the oss-fuzz issues list for details.
Bug fixes:
See above note about oss-fuzz fixes.
CINEON: Fix unexpected hang on a crafted Cineon image. SourceForge issue 571.
Drawing recursion is limited to 100 and may be tuned via the MAX_DRAWIMAGE_RECURSION pre-processor definition.
Fix reading MIFF files using legacy keyword 'color-profile' for ICC color profile as was used by ImageMagick 4.2.9.
Fix reading/writing files when 'magick' is specified in lower case. This bug was a regression in 1.3.30.
New Features:
TIFF: Support Zstd compression in TIFF. This requires libtiff 4.0.10 or later.
TIFF: Support WebP compression in TIFF. This requires libtiff 4.0.10 or later.
API Updates:
MagickMonitor() is marked as deprecated. Code should not be using this function any more.
Feature improvements:
The progress monitor callbacks (registered using MagickMonitor() or MagickMonitorFormatted()) are serialized via a common semaphore rather than via critical sections in OpenMP loops. OpenMP loops are updated to use OpenMP 'atomic' and 'flush' to update shared loop variables rather than using a OpenMP 'critical' construct, reducing contention. Performance on some targets is observed to have been improved by this change.
Build Changes:
There was already a 'compare' command installed with the '--enable-magick-compat' configure option was used but it did not function. Now it functions. There was no compare command in ImageMagick 5.5.2 and this compare command is only roughly similar to a compare command in some subsequent ImageMagick release.
Removed Remove Ghostscript library support (--with-gslib) from configure script. The 'HasGS' pre-processor defines which were enabled by this remain in the source code so it is still possible to use this library if absolutely necessary (e.g. CPPFLAGS=-DHasGS LIBS=-lgs).
No longer explicitly link with the OpenMP library when it will be supplied already due to CFLAGS.
Behavior Changes:
JPEG: Libjpeg-turbo is allowed 1/5th the memory resource limit provided for Graphicsmagick via the cinfo->mem->max_memory_to_use option, which is part of the IJG JPEG API/ABI, but usually not supported there. This feature works for libjpeg-turbo 1.5.2 and later. Limiting the memory usage is useful since libjpeg-turbo may otherwise consume arbitrary amounts of memory even before Graphicsmagick is informed of the image dimensions.
JPEG: The maximum number of JPEG progressive scans is limited to 50. Otherwise some technically valid files could be read for almost forever.
Papirus is a free and open source SVG icon theme, based on Paper Icon
Set with a lot of new icons and a few extras, like Hardcode-Tray support,
KDE colorscheme support, Folder Color support, and others.
The Papirus icon theme is available in four variants:
* Papirus
* Papirus Dark
* Papirus Light
* ePapirus (for elementary OS and Pantheon Desktop)
Package imaging provides basic image processing functions (resize,
rotate, crop, brightness/contrast adjustments, etc.).
All the image processing functions provided by the package accept
any image type that implements image.Image interface as an input,
and return a new image of *image.NRGBA type (32bit RGBA colors,
not premultiplied by alpha).
* Several security bug fixes with malicious GIFs.
* Kill a use-after-free error reported by @junxzm1990.
* Add SIMD support for resizing. When enabled this improves resize
performance enormously for complex resize methods.
* Add thread support for resizing. `-j[NTHREADS]` tells gifsicle to use
up to NTHREADS threads to resize an input image. There are several
caveats---multiple threads can be only used on unoptimized images.
Thanks to Github user @wilkesybear.
* Quashed several crashes and undefined behaviors. Thanks to Github users
including @pornel, @strazzere, and @b0b0505.
* Minor bug fixes.
Release 1.16.0 (2018-10-19 Bryce Harrington <bryce@bryceharrington.org>)
========================================================================
This new stable release incorporates a number of improvements made in
the four years since 1.14.0.
Of particular note is a wealth of work by Adrian Johnson to enhance PDF
functionality, including restoring support for MacOSX 10.4, metadata,
hyperlinks, and more.
Much attention also went into fonts, including new colored emoji glyph
support, variable fonts, and fixes for various font idiosyncrasies.
Other noteworthy changes include GLESv3 support for the cairo_gl
backend, tracking of SVG units in generated SVG documents, and cleanups
for numerous test failures and related issues in the PDF and Postscript
backends.
For a complete log of changes, please see
https://cairographics.org/releases/ChangeLog.1.16.0
Features and Enhancements
-------------------------
* Add support for OpenGL ES 3.0 to the gl backend.
* The PDF backend has gained support for a range of widely used
features, including thumbnails, page labels, metadata, document
outlines, structured text, hyperlinks, and tags. Tags permit adding
logical info such as headings, tables, figures, etc. that facilitates
indexing, accessibility, text reflow, searching, and extraction of the
tagged items to other software. For details on this new PDF
functionality, see:
https://lists.cairographics.org/archives/cairo/2016-June/027427.html
* Variable font support. Variable fonts are single font files with
various typography characteristics, such as weight or slant, that users
of the font can adjust between two points. Effectively this enables a
single font to behave as multiple fonts.
* Restore MacOSX 10.4 support. Cairo had dropped 10.4 support when
moving to the CoreText API. Now we automatically detect which API to
use via dynamic linking, so can resume supporting this older version
of MacOSX.
* Support colored emoji glyphs, stored as PNG images in OpenType fonts.
* Skia backend is removed
* Use Reusable streams for forms in Level 3 Postscript.
* Add CAIRO_MIME_TYPE_EPS mime type for embedding EPS files.
* Add CCITT_FAX mime type for PDF and PS surfaces
* svg: add a new function to specify the SVG document unit
(Bug #90166)
* Use UTF-8 filenames on Windows
API Changes
-----------
Several new APIs were added. No existing APIs were altered.
New PDF functionality:
* cairo_pdf_surface_add_outline
* cairo_pdf_surface_set_metadata
* cairo_pdf_surface_set_page_label
* cairo_pdf_surface_set_thumbnail_size
* cairo_tag_begin
* cairo_tag_end
* CAIRO_STATUS_TAG_ERROR
New error status items for problems relating to PDF tagging:
* CAIRO_STATUS_WIN32_GDI_ERROR
* CAIRO_STATUS_FREETYPE_ERROR
* CAIRO_STATUS_PNG_ERROR
New error status items for handling of GDI, libfreetype, and libpng
errors, respectively.
Setting up Win32 surfaces for HDC with alpha channels:
* cairo_win32_surface_create_with_format
New API for added PDF functionality (see above), and new error
status item for problems relating to PDF tagging.
Variable fonts:
* cairo_font_options_get_variations
* cairo_font_options_set_variations
Tracking units in SVG documents:
* cairo_svg_surface_set_document_unit
* cairo_svg_surface_get_document_unit
Dependency Changes
------------------
None
Performance Optimizations
-------------------------
None
Notable Bug Fixes
-----------------
* Fix thin lines that don't show up when printing in Inkscape due to
overly aggressive culling.
(Bug #77298)
* Fix playback of recording surfaces into PDF surfaces, where objects
with negative coordinates were not getting drawn. To address this,
the coordinate systems for PDF and PS have been changed to match
cairo's coordinate system. This allows recording surfaces to be
emitted in cairo coordinates, and results in the same origin being
used for all operations when using the recording surface XObject.
Test cases for PDF and PS have also been updated accordingly.
(Bug #89232)
* Fix "invalidfont" error on some printers when printing PDFs with
embedded fonts that have glyphs (such as spaces) with
num_contours == 0. (Bug #79897)
* Fix missing glyphs such as thin dashes, which get scaled to 0 in
userspace and thus have their drawing operations culled. (Bug #94615)
* Fix other oddities caused by variously idiosyncratic fonts.
* Fix a data race in freed_pool discovered by Firefox's cairo usage.
The patch adads atomic int load and store functions, with relaxed
memory ordering. (Bug #90318)
* Handle SOURCE and CLEAR operators when painting color glyphs.
(Bug #102661)
* Fix falling back to system font with PDFs using certain embedded
fonts, due to truncated font names.
(Bug #103249)
* Prevent curved strokes in small ctms from being culled from vector
surfaces
(Bug #103071)
* Fix assertion hit with PDFs using Type 4 fonts rendered with user
fonts, due to error when destroying glyph page.
(Bug #103335)
* Prevent invalid ptr access for > 4GB images.
(Bug #98165)
* pdf: Fix internal links pointing to other pages, by pre-calculating
page heights so that link positions can be calculated more accurately.
* Fix error reporting in the xcb backend if fallback fails. Instead of
returning NULL when the X11 server can't do some operation, return a
surface in an error state.
* Clarify documentation regarding device scale inheritance and the units
used in cairo_surface_create_similar_image.
(Bug #99094)
* Call XSync in the xlib backend before setting the error handler to
ignore errors for certain requests, to make sure all pending errors
are handled first.
* Fix regression with text containing space character on Win32.
(Bug: https://gitlab.freedesktop.org/cairo/cairo/issues/339)
For a more comprehensive listing of fixed bugs, see the release notes for the
individual 1.15.x releases.
Release 1.15.14 (2018-09-19 Bryce Harrington <bryce@bryceharrington.org>)
============================================================================
We're nearly ready to finalize the 1.16.0 release, so this snapshot
can be considered a beta for 1.16.
The most notable change this release is a performance optimization for
windows, discussed below. Other than that, much of the development
focus was on final polish and stability as we prepare for 1.16.
Some attention went into getting the testsuite passing at least for the
image backend. The Cairo testsuite depends on external software like
Pixman, and changes in the rendering behavior of these dependencies
change test behavior, leading to false positives.
Results from the Coverity static testing tool were also reviewed. Most
of the issues flagged were false positives, but there were several
legitimate problems found and fixed.
For a complete log of changes, please see
https://cairographics.org/releases/ChangeLog.1.15.12
Features and Enhancements
-------------------------
* Add more FreeeType font color conversions to support COLR/CPAL
* Update test reference images against current pixman
API Changes
-----------
None
Dependency Changes
------------------
None
Performance Optimizations
-------------------------
Vasily Galkin introduced a Win32 performance optimization for
CAIRO_OPERATOR_SOURCE when copying data from a backbuffer to an argb32
surface corresponding to a Win32 DC. With this, argb32 drawing should
perform as fast as typical dibsection-buffered GDI drawing. See the
Cairo mailing list for April 2018 for data and discussion of the
performance improvements.
Bug Fixes
---------
* Fix crash when rendering Microsoft's Segoe UI Emoji Regular font.
* Fix build breakage with glesv3 enabled due to non-existant glesv3.pc.
* Fix memory leaks found by Coverity
* Fix incorrect null ptr handling found by Coverity
* Fix test compilation when font-config is disabled
* Use _cairo_malloc instead of malloc (Bug #101547) (CVE-2017-9814)
* Fix assertion failure in the freetype backend (Bug #105746)
Release 1.15.12 (2018-04-04 Bryce Harrington <bryce@osg.samsung.com>)
========================================================================
The main focus for this release is the addition of Variable Font
support. Variable fonts are single font files with various typography
characteristics, such as weight or slant, that users of the font can
adjust between two points. Effectively this enables a single font to
behave as multiple fonts.
The Skia backend is disabled in this release, due to severe bitrot, and
will be removed in future releases. Contact the cairo team if you have
a need of this backend.
For a complete log of changes, please see
https://cairographics.org/releases/ChangeLog.1.15.12
Features and Enhancements
-------------------------
* Variable font support
* Skia backend is disabled
API Changes
-----------
* cairo_font_options_get_variations() and
cairo_font_options_set_variations() are added.
Dependency Changes
------------------
None
Performance Optimizations
-------------------------
None
Bug Fixes
---------
* Fix errors in csi-trace --help and --version options
* Fix a 'memory leak' in the image compositor, with
pixman_glyph_cache_t.
* Fix access of uninitialized memory found by valgrind
(Bug #91271)
* Fix improper initialization of memory in
_cairo_ft_font_face_create_for_pattern()
(Bug #105084)
* Fix multi-monitor virtual desktop with negative coords on Win32
(Bug #100793)
* Fix issues occuring with older FreeType versions.
Release 1.15.10 (2017-12-07 Bryce Harrington <bryce@osg.samsung.com>)
========================================================================
This release adds GLESv3 support to the cairo_gl backend, adds
tracking of SVG units in generated svg documents, and cleans up numerous
test failures and related issues in the PDF and Postscript backends.
For a complete log of changes, please see
https://cairographics.org/releases/ChangeLog.1.15.10
Features and Enhancements
-------------------------
* Add support for OpenGL ES 3.0 to the gl backend.
* Use Reusable streams for forms in Level 3 Postscript.
* Add CAIRO_MIME_TYPE_EPS mime type for embedding EPS files.
* Add CCITT_FAX mime type for PDF and PS surfaces
* svg: add a new function to specify the SVG document unit
(Bug #90166)
* Use UTF-8 filenames on Windows
API Changes
-----------
* cairo_svg_surface_set_document_unit() and
cairo_svg_surface_get_document_unit()
Dependency Changes
------------------
None
Performance Optimizations
-------------------------
None
Bug Fixes
---------
* Fix regression in gles version detection
* Fix undefined-behavior with integer math.
* Handle SOURCE and CLEAR operators when painting color glyphs.
(Bug #102661)
* Convert images to rgba or a8 formats when uploading with GLESv2
* Use _WIN32 instead of windows.h to check for windows build.
* Fix sigabrt printing documents with fonts lacking the mandatory .nodef
glyph.
(Bug #102922)
* Prevent curved strokes in small ctms from being culled from vector
surfaces
(Bug #103071)
* Fix painting an unbounded recording surface with the SVG backend.
* Fix falling back to system font with PDFs using certain embedded
fonts, due to truncated font names.
(Bug #103249)
* Fix handling of truetype fonts with excessively long font names
(Bug #103249)
* Fix race conditions with cairo_mask_compositor_t
(Bug #103037)
* Fix build error with util/font-view
* Fix assertion hit with PDFs using Type 4 fonts rendered with user
fonts, due to error when destroying glyph page.
(Bug #103335)
* Set default creation date for PDFs
* Prevent invalid ptr access for > 4GB images.
(Bug #98165)
* Prevent self-copy infinite loop in Postscript surface.
* Fix padded image crash in Postscript surface.
* Fix annotation bugs in PDFs and related memory leaks
* Fix test failures and other assorted issues in ps and pdf code.
* Fix code generation when using GCC legacy atomic operations
(Bug #103559)
* Fix various compilation warnings and errors.
* Fix various distcheck errors with private symbols, doxygen formatting,
etc.
Release 1.15.8 (2017-08-29 Bryce Harrington <bryce@osg.samsung.com>)
========================================================================
This small snapshot provides new colored emoji glyph support, and a
handful of minor fixes.
For a complete log of changes, please see
https://cairographics.org/releases/ChangeLog.1.15.8
Features and Enhancements
-------------------------
* Support colored emoji glyphs, stored as PNG images in OpenType fonts.
API Changes
-----------
None
Dependency Changes
------------------
None
Performance Optimizations
-------------------------
None
Bug Fixes
---------
* pdf: Fix internal links pointing to other pages, by pre-calculating
page heights so that link positions can be calculated more accurately.
* image: Fix crash on negative lengths
* win32: Fix initialization of mutexes for static builds
* pdf: Don't emit /PageLabel dict when no labels defined
* font: Fix color font loading on big-endian systems
* font: Fix color font support infinite-loop with empty glyphs
* Fix off by one check in cairo-image-info.c
Release 1.15.6 (2017-06-13 Bryce Harrington <bryce@osg.samsung.com>)
========================================================================
This new snapshot incorporates changes over the past half-year since the
1.15.4 snapshot, including all the fixes from the 1.14 release series.
The PDF code continues to be enhanced, and we're restored MacOSX 10.4
support. Font-related fixes and improved error handling for X round out
the release.
For a complete log of changes, please see
https://cairographics.org/releases/ChangeLog.1.15.6
Features and Enhancements
-------------------------
* Detect if variable fonts have synthesized bold/italic or non-default
variants, and use a fallback font where needed.
* Restore MacOSX 10.4 support. Cairo had dropped 10.4 support when
moving to the CoreText API. Now we automatically detect which API to
use via dynamic linking, so can resume supporting this older version
of MacOSX.
API Changes
-----------
None
Dependency Changes
------------------
None
Performance Optimizations
-------------------------
None
Bug Fixes
---------
* Fix error reporting in the xcb backend if fallback fails. Instead of
returning NULL when the X11 server can't do some operation, return a
surface in an error state.
* Call XSync in the xlib backend before setting the error handler to
ignore errors for certain requests, to make sure all pending errors
are handled first.
* Fix text-glyph-range for quartz-font. Use 0xFFFF instead of 0 for
invalid index tracking.
* Fix handling of Supplementary Multilingual Plane (SMP) Unicode
characters in quartz-font.
* Fix various issues in the drm backend including updating API usage and
general code cleanup.
* Clarify documentation regarding device scale inheritance and the units
used in cairo_surface_create_similar_image.
Bug #99094.
Release 1.15.4 (2016-12-9 Bryce Harrington <bryce@osg.samsung.com>)
=======================================================================
This new snapshot incorporates changes over the past year since the
1.15.2 snapshot, including all the fixes from the 1.14 release series.
Of particular note in this snapshot is a wealth of work by Adrian
Johnson to enhance PDF support, as well as numerous bug fixes provided
by him and other contributors.
For a complete log of changes since the last release, please see:
https://cairographics.org/releases/ChangeLog.1.15.4
Features
--------
* The PDF backend has gained support for a range of widely used
features, including thumbnails, page labels, metadata, document
outlines, structured text, hyperlinks, and tags. Tags permit adding
logical info such as headings, tables, figures, etc. that facilitates
indexing, accessibility, text reflow, searching, and extraction of the
tagged items to other software. For details on this new PDF
functionality, see:
https://lists.cairographics.org/archives/cairo/2016-June/027427.html
API Changes
-----------
cairo_win32_surface_create_with_format
Added a cairo API to set up Win32 surfaces for HDC with alpha channels.
cairo_pdf_surface_add_outline
cairo_pdf_surface_set_metadata
cairo_pdf_surface_set_page_label
cairo_pdf_surface_set_thumbnail_size
cairo_tag_begin
cairo_tag_end
CAIRO_STATUS_TAG_ERROR
New API for added PDF functionality (see above), and new error
status item for problems relating to PDF tagging.
CAIRO_STATUS_WIN32_GDI_ERROR
CAIRO_STATUS_FREETYPE_ERROR
CAIRO_STATUS_PNG_ERROR
New error status items for handling of GDI, libfreetype, and libpng
errors, respectively.
Dependency Changes
------------------
None
Performance Optimizations
-------------------------
None
Bug Fixes
---------
* Bug fixes from 1.15.2 (see the 1.15.2 NEWS for details)
* Fix playback of recording surfaces into PDF surfaces, where objects
with negative coordinates were not getting drawn. To address this,
the coordinate systems for PDF and PS have been changed to match
cairo's coordinate system. This allows recording surfaces to be
emitted in cairo coordinates, and results in the same origin being
used for all operations when using the recording surface XObject.
Test cases for PDF and PS have also been updated accordingly.
(Bug #89232)
* Fix "invalidfont" error on some printers when printing PDFs with
embedded fonts that have glyphs (such as spaces) with
num_contours == 0. (Bug #79897)
* Fix missing glyphs such as thin dashes, which get scaled to 0 in
userspace and thus have their drawing operations culled. (Bug #94615)
* Fix other oddities caused by variously idiosyncratic fonts.
* Fix deadlock when destruction of a scaled font indirectly triggers
destruction of a second scaled font, causing the global cache to be
locked twice. (Bug #93891)
* Fix X errors reported to applications when shmdt() is called before
the Attach request is processed, due to missing xcb and xlib calls.
* Fix random failure in record-paint-alpha-clip-mast test case, caused
by an incorrect assumption that a deferred clear can be skipped.
(Bug #84330)
* Fix crash when dealing with an XShmGetImage() failure, caused by a
double free in _get_image_surface(). (Bug #91967)
* Fix invalid execution of ASCII85 data by the PS interpreter that the
image operator didn't use, by flushing the extraneous data after
drawing the image. (Bug #84811)
* Fix decoding of Adobe Photoshop's inverted CMYK JPEG files in PDF
export.
* Fix unbounded surface assertion in win32-print code.
* Fix a data race in freed_pool discovered by Firefox's cairo usage.
The patch adads atomic int load and store functions, with relaxed
memory ordering. (Bug #90318)
* Cleanup debugging text sent to stdout instead of log. (Bug #95227)
* Fix build issue when using non-GNU strings utility. (Bug #88639)
* Fix build of cairo modules as regular modules, not as versioned shared
libaries. (Bug #29319)
* Fix build on win32 using gcc 5.4.
* Fix build of script backend to require zlib.
* Update test suite reference images using Debian Jessie 64-bit and
poppler current as of June, 2016.
* Various improvements to documentation and tests, compiler warning
fixes, and an assortment of code refactoring and cleanup.
Release 1.15.2 (2015-12-10 Bryce Harrington <bryce@osg.samsung.com>)
========================================================================
This release is largely a rollup to include a variety of fixes that
didn't make the cut for the stable 1.14.2 and 1.14.4 releases, as well
as all the fixes from those releases. Notably this includes a highly
requested new API for Win32 surfaces.
For a complete log of changes since the last release, please see:
https://cairographics.org/releases/ChangeLog.1.15.2
Features
--------
None
API Changes
-----------
cairo_win32_surface_create_with_format
Added a cairo API to set up Win32 surfaces for HDC with alpha channels.
Dependency Changes
------------------
None
Performance Optimizations
-------------------------
None
Bug Fixes
---------
* All the bug fixes from 1.14.2, 1.14.4, and 1.14.6
* Fix xcb/xlib compilation and calls. Make image boxes behave when SHM
is not available.
* Fix various issues with printing and transparent images on Win32.
* Fix thin lines that don't show up when printing in Inkscape due to
overly aggressive culling.
(Bug #77298)
* Fix broken printing via pdf when glyph 0 is used for rendering,
resulting in missing spaces and letters.
(Bug #89082)
* Fix crash for certain glyphs in opentype fonts.
(Bug #91902)
* Fix incorrect rendering of SVG paths with more than one subpath. If
more than one trap is passed in then it's guaranteed that the returned
traps will have their left edge to the left of their right edge, but
if only one trap is passed in then the function always returns without
doing anything.
(Bug #90984)
* Improve rendering with Quartz to better match pixman's blending and
filtering behavior.
Release 1.14.6 (2015-12-09 Bryce Harrington <bryce@osg.samsung.com>)
========================================================================
Simple bugfix release to fix one Windows issue.
For a complete log of changes since 1.14.4, please see:
https://cairographics.org/releases/ChangeLog.1.14.6
Features
--------
None
API Changes
-----------
None
Dependency Changes
------------------
None
Performance Optimizations
-------------------------
None
Bug Fixes
---------
* Fix failure on Windows due to reference of the function
cairo_win32_surface_create_with_format(), which isn't included in the
1.14.4 release. (Bug #92771)
Release 1.14.4 (2015-10-28 Bryce Harrington <bryce@osg.samsung.com>)
========================================================================
Just in time for Halloween we see another bug-fix release for Cairo.
This brings a few dozen straightforward bug fixes with no API changes.
In addition, this includes a typical assortment of fixes to tests,
cleanup of warnings and memory leaks, correction of misspellings,
updates to documentation, etc.
For a complete log of changes since 1.14.2, please see:
https://cairographics.org/releases/ChangeLog.cairo-1.14.4
Features
--------
None
API Changes
-----------
None
Dependency Changes
------------------
None
Performance Optimizations
-------------------------
None
Bug Fixes
---------
* Avoid appending empty slots to user data arrays. Fixes a memory
consumption regression since commit 9341c254a.
* Return a better error (file-not-found) when setting up pango on
devices where the font files don't have read permissions.
* Fix regression in the font size of canvas text in Inkscape when
compiled with the Quartz backend. (Bug #84324)
* Fix _cairo_gl_shader_bind_matrix() to maintain compatibility with
OpenGL ES 2.0. Manually transpose the matrix.
* Fix incorrect font descriptor conversion when the font matrix yy is
negative. (Bug #90538)
* Fix crash when using a complex path for clip and stroke due to
discarding the intersection exactly at the top edge.
(Bug #74779)
* Fix cairo_get_locale_decimal_point() on Android
* Fix compilation problem on AIX due to conflicting usage of symbol
'jmpbuf'. (Bug #89339)
* Fix broken rendering with XCB due to snapshotting of uploaded part of
surfaces. (Bug #67505)
* Fix loss of alpha when copying a mask for a cairo recording surface,
resulting in a double copy. (Bugs #73038, #73901)
* Fix incorrect recording of certain paths with script surfaces.
(Bug #91054)
* Fix typo in definition of MAYBE_WARN in configure script.
(Bug #89750)
* Fix use of filename variable after it's been freed.
(Bug #91206)
* Fix out of bounds access when printing pattern.
(Bug #91266)
* Fix incorrect size calculation in glyph cache unlocking for Cairo GL
compositor.
(Bug #91321)
* Fix memory leak in _cairo_gl_pattern_texture_setup()
(Bug #91537)
* Fix transparent images in win32-print.
(Bug #91835)
* Fix _put_shm_image_boxes and _put_image_boxes when no SHM available
with XCB.
It has been a year since the previous release. This is the first
release made from the Git repository at
https://gitlab.com/libtiff/libtiff using a collaborative process.
Since the previous release, a number of security issues have been
fixed, and some significant new features have been added.
This release adds support for Zstd and WebP compression algorithms.
In their own way, each of these compression algorithms is highly
complimentary to TIFF.
Zstd provides improved compression and decompression speed vs zlib's
Deflate as well as a broader range of compression ratios. Zstd is
developed by Facebook and the implementation continues to be improved.
WebP is optimized for small/medium 8-bit images while offering
improved compression performance vs traditional JPEG. WebP works well
in strips or tiles to compress large images down to very small files,
while preserving a good looking image. WebP is developed by Google,
and its implementation continues to be improved.
Due to Adobe's TIFF tag registration interface going off-line, we have
had to assign our own tags for Zstd and WebP.
v1.15 2017/08/05
(2017/08/05) PS1 - use round() instead of rint()
(2017/08/05) PS1 - portability: set both _POSIX_C_SOURCE and
_NETBSD_SOURCE to enable strdup prototype in string.h, as
_XOPEN_SOURCE=500 doesn't work on macOS, _XOPEN_SOURCE=600 doesn't
work on SunOS, and _POSIX_C_SOURCE alone doesn't work on Minix.
(2017/08/02) PS1 - fixed a potential arithmetic overflow.
(2017/06/17) PS1 - fixed a bug with very large bitmaps. Reported
by Michael Voříšek.
(2017/03/05) PS1 - autoconf: added --enable-local-getopt to avoid
using the system-wide getopt library (it may still end up being
used on GNU systems).
(2017/03/05) PS1 - portability: moved getopt.h to a subdirectory,
to be included only when needed. Reported by Daniel Macks.
(2017/03/05) PS1 - portability: use _XOPEN_SOURCE=600 to fix macOS
warning.
(2017/02/27) PS1 - fixed more buffer overflows (not security
critical), including one reported by Agostino Sarubbo. This also
fixes CVE-2017-12067.
V4.0.0:
New OCR engine
- Added a new OCR engine that uses neural network system based on LSTMs, with major accuracy gains.
- This includes new training tools for the LSTM OCR engine. A new model can be trained from scratch or by fine tuning an existing model.
- Added trained data that includes LSTM models to 123 languages.
- Added optional accelerated code paths for the LSTM recognizer:
* Using OpenMP
* Using SIMD: AVX2 / AVX / SSE4.1
- Added a new parameter lstm_choice_mode that allows to include alternative symbol choices in the hOCR output.
- The new LSTM engine still does not support all features from the old legacy engine (see missing features).
Other OCR engines
- The pattern matching OCR engine that was the primary OCR engine in previous versions is still available in this version.
- Removed the 'Cube' OCR engine from the codebase. It was used for Hindi and for Arabic. The New LSTM engine performs much better, thus the Cube engine was no longer needed.
Updated build system
- Tesseract now uses semantic versioning.
- Tesseract now requires Leptonica 1.74.0 or a higher version.
- For building Tesseract from source code, a compiler with good C++ 11 support is required. See here for a list of officially supported compilers.
- Added unit tests to the main repo. The unit tests require Git submodules and the code for training.
- Added an option to compile Tesseract without the code of the legacy OCR engine.
- Update minimum required autoconf version to 2.63.
- Training tools dependencies - Update minimum required versions: ICU 52.1, Pango 1.22.0.
- Reorganized Tesseract's source tree. Most sources are now below the src directory.
Bug fixes and enhancements
- Fixed many issues that triggered compiler warnings.
- Fixed many issues reported by Coverity Scan or LGTM.
- Fixes to trainingdata rendering.
- Fixed damage to binary images when processing PDFs.
- Don't trigger a deliberate segmentation fault for fatal errors in release code.
- Fixed some issues in OpenCL code. OpenCL now works for the legacy Tesseract OCR engine, but does not improve the performance. It is not implemented for the LSTM OCR engine.
- Improved multi-page TIFF handling.
- Improvements to PDF rendering.
- Added version information and improved help texts to the training tools.
- Added faster version of log2().
- Documented in tesseract man page the option to use an input text file which contains lists of images.
- Made 'osd' the default traineddata when psm 0 is requested (currently this feature is only implemented in the command line interface, but not in the API).
- Removed tessedit_pageseg_mode 1 from hocr, pdf, and tsv config files. The user should explicitly use --psm 1 if that is desired.
- The list of available languages and scripts is now sorted alphabetically.
- Parameter unlv_tilde_crunching changed to false, because of default values cause issues in cases of unlv output in Tesseract 4.
- Removed obsolete code.
Upstream changes (from NEWS):
== Ruby-GNOME2 3.3.0: 2018-10-31
This is a release for fixing memory leak bugs of cairo-gobject,
improving macOS support and adding support for GEGL.
=== Changes
==== Ruby/GLib2
* Improvements
* Added support for the same constant name and class name for error.
* Fixes
* Fixed wrong constant values:
* (({GLib2::MINLONG}))
* (({GLib2::MAXLONG}))
* (({GLib2::MAXULONG}))
* (({GLib2::MINUINT64}))
* (({GLib2::MAXSIZE}))
* (({GLib2::MINFLOAT}))
* (({GLib2::MINDOUBLE}))
[GitHub#1244][Reported by cobodo]
==== Ruby/CairoGObject
* Fixes
* Fixed a memory leak.
[GitHub#1232][Reported by Jean-Christophe Le Lann]
* Stopped to increase needless reference.
[GitHub#1079][Reported by noanoa07]
==== Ruby/GObjectIntrospection
* Improvements
* Added support for transfer full output parameter.
* Fixes
* Fixed a bug that class method closure doesn't work.
[GitHub#1245][Reported by kojix2]
==== Ruby/GdkPixbuf2
* Improvements
* (({GdkPixbuf::Pixbuf.new})): Added support for auto row stride
detection for (({[Integer]})) data.
==== Ruby/Pango
* Improvements
* Made test more robust.
[GitHub#1239][Reported by Michael Hudson-Doyle]
==== Ruby/GTK3
* Improvements
* Improved backward compatibility for (({Gtk::ListStore#set_column_types})).
[GitHub#1240][Reported by Edward Hennessy]
* Fixes
* Fixed wrong size used bug on HiDPI.
[GitHub#1079][Reported by noanoa07]
==== Ruby/Poppler
* Improvements
* Added support for Popper 0.70.0.
==== Ruby/GEGL
* Improvements
* Added.
=== Thanks
* Jean-Christophe Le Lann
* Michael Hudson-Doyle
* Edward Hennessy
* cobodo
* kojix2
* noanoa07
From 681748ec2f5ce88da5f9fa6831e1653e46af8a66 Mon Sep 17 00:00:00 2001
From: Even Rouault <even.rouault@spatialys.com>
Date: Sun, 14 Oct 2018 16:38:29 +0200
Subject: [PATCH 1/1] JBIG: fix potential out-of-bounds write in JBIGDecode()
JBIGDecode doesn't check if the user provided buffer is large enough
to store the JBIG decoded image, which can potentially cause out-of-bounds
write in the buffer.
This issue was reported and analyzed by Thomas Dullien.
Also fixes a (harmless) potential use of uninitialized memory when
tif->tif_rawsize > tif->tif_rawcc
And in case libtiff is compiled with CHUNKY_STRIP_READ_SUPPORT, make sure
that whole strip data is provided to JBIGDecode()
The last part (CHUNKY_STRIP_READ_SUPPORT) was adapted by myself to fit
the libtiff release.
Bump PKGREVISION.
