[gegl] Created tag GEGL_0_1_2

[Martin Nordholts <martinn src gnome org>]

The signed tag 'GEGL_0_1_2' was created.

Tagger: Martin Nordholts <martinn src gnome org>
Date: Sun Oct 31 22:38:19 2010 +0100

    Tag for the GEGL 0.1.2 release

Changes since the last tag 'GEGL_0_1_0_REAL':

Adam Turcotte (3):
      Bug 552658 â?? redundant/incorrect code in gegl-buffer.c (check for ...
      Improved version of the Nohalobox samplers (downsize, downsharp, and downsmooth). A small improvement for each sampler is being developed, and downsmooth will be re-written. This patch fixes some bugs that were re-instated by overriding a previous commit in the previous patch.
      Add new samplers from gsoc

Alexia Death (5):
      Add mirrors op to workshop that implements GIMP kaleidoskope pluging for gegl
      Fixes to the mirrors op
      Making the mirrors op trim handles act saner
      Changes to mirrors that fix some bugs in warping and adds more controll over output size
      Further fixes and a new option, zoom, for mirrors op.

Danny Robson (2):
      tests: Compare change-rect test buffers with small epsilon
      Check cached parameters to avoid relinking nodes

Debarshi Ray (1):
      affine: Add fast paths for axis aligned reflect

Javier Jardón (1):
      Bug 589769 - Remove deprecated Glib and GTK+ symbols

Martin Nordholts (30):
      Bump version to 0.1.1
      Remove ChangeLog from version control
      Add Makefile.am rule to generate ChangeLog
      Add ChangeLog.pre-0-1
      Enable automake silent-rules by default
      Incorporate minor version in soname
      Remove invalid comment in configure.ac about soname
      Bug 589124 â?? gobj2dot.rb missing from gegl tarball
      operations: Only build the display op on UNIX platforms
      gegl: Remove compiler warning about copying uninitialized memory
      gegl: Properly build module_path for win32
      Revert "gegl: Properly build module_path for win32"
      operations: #define SIGQUIT 3 in display.c on win32 to fix compilation
      Add $(EXEEXT) to tools in Makefile.ams to fix win32 compilation
      Revert "operations: Only build the display op on UNIX platforms"
      Update .gitignore for win32
      operations: Consistently use dash as word separator in ops
      examples: Update .gitignore
      examples: #include <glib/gstdio.h> in float-lookup.c
      operations: Fix cast in nop.c
      operations: Fix warning about mixed code and variables in path.c
      operations: Namespace functions in "gegl:jpg-load"
      gegl: Clarify that gegl_processor_destroy() will always only unref
      docs: Put inheritance.png in builddir
      operations: Namespace static functions for "gegl:png-save"
      docs: Add paragraph about useful functions for interactive debugging
      gegl: Simplify gegl_have_visitor_visit_node()
      gegl: Add have_rect debug output
      gegl: Format gegl_buffer_interpolation_from_string() slightly
      tests: Add regression test for simple scaling

Michael Natterer (2):
      Use g_slice_new0() for GeglLookup so the tables don't contain garbage
      Pull the pointer into the GeglLookupFunction typedef; formatting

Michael Schumacher (3):
      Re-enabling the display op.
      Bug 588172 - libgegl uses deprecated g_win32_get_package_installation_subdirectory ()
      Fix typo: G_OS_WIN_32 -> G_OS_WIN32

Nils Philippsen (2):
      Bug 587684 â?? building gegl fails in operations/workshop/external
      use ARGV.length instead of ARGV.count

Ruben Vermeersch (6):
      Fix Gegl# build.
      Auto-generate ChangeLog files for Gegl#.
      Moved configure.in to configure.ac.
      Add a piece of hand-crafted Babl API, to provide Babl#.
      Expose GeglBuffer in Gegl#.
      Bring version number in sync with main GEGL.

Sven Neumann (1):
      require automake >= 1.9.6 and libtool >= 1.5

Yaakov Selkowitz (2):
      Bug 596229 -  gegl Cygwin patches : no-undefined patch
      Bug 596229 -  gegl Cygwin patches : VPATH patch

Ã?tienne BERSAC (1):
      Correct my forname :x

�yvind Kolås (107):
      gegl-xml: specify a gegl:nop op for top level element
      Added an op that can be used to fix a wrongly premultiplied buffer.
      operation-point-composer special case fast path for opacity
      Removed fast paths from gegl-operation-point-composer.c
      Made gegl_buffer_get_abyss public.
      Made gegl_operation_context_get_object public
      Added fast path to a stand-alone over op itself
      Add fast path to pass buffer through opacity op
      Removed dead code
      Do not create a sub-buffer in nop, but pass the original straight through.
      Fix bug in gegl_buffer_linear_open
      Do not rely on input or output buffer extents in ops,     removing the need for GEGL to pass sub-buffers with     the input/output bounding boxes.
      Build ops with external dependencies last.
      Share an empty buffer
      Made normal be a custom op with the same shortcuts as over
      gaussian-blur, fix assumption about extent of output buffer.
      Provide direct access to the cache buffer, do not use a sub buffer
      Made point filter and composers able to operate in-place.
      Fix build, the previous commit was incomplete.
      Remove unneeded ref/unref of interface
      Add a hashtable to the tile cache
      Restore GeglProcessor queued_region member
      Updated gegl-paint code to reflect new property names on the path op.
      Revert commit 5658532 it is causing a buffer leak
      When doing in-place processing use it only once in the iterator
      Improved locking in GEGL
      [buffer] Guard some multi threading code in #if #endif
      Fix typo in multi threading guards (s/ENABLE_MP/ENABLE_MT)
      Made GeglTile be a struct instead of an object
      Disable in-place processing when multi threading is enabled.
      [affine] Avoid using variables in the op for intermediate computations.
      Disable internals of gegl_bt() when on win32
      [tile] Use g_atomic_int_inc and g_atomic_int_dec_and_test for refcount.
      [buffer] add #if ENABLE_MP guards
      Use g_atomic_int_inc to increment the unique number for the swapname
      Enabled in-place processing for point filters/composers
      Made the number of threads configurable.
      Removed unused hack for output proxy pads
      Made gegl_node_get_producer() return the graph node of meta-ops
      Cleaned up API comment string inconsistencies.
      Added syntactic sugar for graph construction for C
      Trimmed down ang reindented csugar example
      Removed babl processing time instrumentation
      Replace g_object_unref with gegl_tile_unref a couple of places.
      [op:grey] Add a dummy value copying loop to process()
      Made gegl_node_get_depends on static
      Removed some typechecks in performance critical internal API
      Fix Makefile installed to ~/.local/share/gegl-0.1/plug-ins
      Moved specilisation logic from base class.
      Added an implementation of posterize to workshop
      Require glib 2.22.0 or newer.
      Made the stamp mask for stroking the path static.
      never pass through
      [buffer] Use slice allocator for GeglBufferIterator
      Re-enable shared read-write buffers.
      Fix pool in buffer iterator
      Fixed mistake made when refactoring logic away from base class
      Reduce number of needed iterators.
      Remove reference to non-existing file in Makefile.am
      Added a SIMD version of gegl:opacity
      Fix bug in previous commit.
      gegl:opacity Only build process_simd() when HAS_G4FLOAT is defined
      buffer: Fix double opening of input stream in file tile backend
      Added general caching floating point lookup table code.
      Moved GeglLookup to its own files.
      GeglOperationPointComposer fixed a typobug
      GeglOperationPointFilter Removed unused variable
      GeglLookup: Made the precision/dimensions configurable
      Removed gegl:normal duplicates the code of gegl:over
      gaussian-blur: do not force FIR filter
      Removed duplicate #define g4float_half
      adjust column layout of time instrumentation
      gegl:layer fix passthrough redirection in prepare when there is no source
      GeglNodePrivate reduced maximum number of evalmgrs to 8
      tools/xml_insert.sh specify bash in #!
      rgegl - updated an example
      gegl_node_invalidated accept NULL as region to invalidate
      gegl:text Invalidated computed data (and caches) when text layout changes.
      gegl:unsharp-mask rearrange graph to allow more in-place processing
      Annotate temporary buffers to restrict in-place processing.
      Use separate iterations indices for read/write when iterating.
      Removed tonemap operation
      Added macros for annotating objects passed in graph.
      gegl:lua fix bitrot in build/code
      configure: added warning that --enable-mt is unstable
      Updated NEWS for 0.1.2 release.
      configure: Require babl 0.1.2 or newer
      gegl_exit: Removed stray g_printf
      Added an XML composition driven test system
      Fix XML composition based regression test system.
      Fix XML composition based regression test system.
      img_cmp: correct computation of % pixels wrong
      Replace reflect reference image with a better one
      img_cmp: report error as Î?e tolerate a max Î?e of 1.5 as no-error
      Update AUTHORS
      Replaced car-stack.jpg in tests/compositions with png
      Remove text from clones test in test/compositions
      img_cmp: Write an debug output file highlighting pixel differences
      buffer: remove gegl_buffer_lock/unlock from gegl_buffer_sample
      ff-load: Clean up deprecation warning from ffmpeg
      GeglSampler: cache a bablfish in the sampler
      gegl_buffer_sample: recreate the sampler if format changes
      gegl_buffer_sample: fix issue in previous commits bugfixing
      Added performance tracking framework
      Add tests and reference images for resamplers
      update NEWS and AUTHORS
      Release 0.1.2