In case of negative y in the region to process, we were accessing invalid memory
(negative array index).
I hesitated between make so that a given ordinate always use the same index or
if we just want the start ordinate (whatever it is) to use index 0. The later
could have just been `(y - result->y) % RANDOM_TABLE_SIZE`.
I just decided to keep the existing logic (former case) though to be fair, not
sure it matters much.
I somewhat bisected GEGL commits between 0.4.20 and 0.4.24 and found
that the one that introduced the env var GEGL_OPERATION_INIT_OUTPUT is
the first showing the problem.
Reviewer (Jehan) note: so it would be commit 6e9610e65c on GEGL repo.
This fix makes sense as it means that since this commit, the output
buffer could have random values. It's not a problem for any operation
where we fill every value, but I guess it's not the case for
"gimp:cage-coef-calc" which was likely relying on the old behavior of
being 0-initialized.
(cherry picked from commit 3b2a8f79f0)
When converting Brightness-Contrast to Levels, and Levels to
Curves, make sure to copy the common settings as well as the
operation-specific settings.
(cherry picked from commit 987447f18f)
In gimp_operation_config_sync_node(), when the operation has a
property of the config object's type, don't skip the other
properties. This makes sure to set the "linear" property of
GimpOperation{Curves,Levels} according to the config object. We'd
previously done it manually in GimpFilterTool, but the setting was
not applied when repeating the filter.
(cherry picked from commit 0096e563f6)
As per commit ed7ea51fb7, reintroduce
the "Fade" functionality for filters, by incorporating it directly
into GimpFilterTool.
Add "mode" and "opacity" options to GimpOperationSettings, and add
a corresponding "Fade" expander to the GimpFilterTool dialog
allowing to control them.
Reintroduce the FADE layer-mode context, and use it to mark the
layer modes avaialable for fading.
Add a new GimpOperationSettings class, to be used as a base class
for all operation-config types. The class provides options common
to all operations (namely, the clipping mode, input region, and
color options), which were previously stored in GimpFilterOptions,
and were therefore bound to the filter tool, instead of being
stored as part of the operation settings; as a result, these
options would have no effect when reapplying a filter, or when
restoring a preset.
The GimpOperationSettings options do not affect the operation
node, but rather the associated GimpDrawableFilter object. The
class provides a gimp_operation_settings_sync_drawable_filter()
function, which applies the options to the filter.
Modify all custom and auto-generated operation-config types to
derive from GimpOperationSettings, and modify the GimpConfig
functions of the former to account for the GimpOperationSettings
properties, using a set of protected functions provided by the
class.
In GimpOperationLayerMode, when the op has a mask connected, and
we're processing an area outside the mask bounds, set the op's
opacity to 0, so that the backdrop shows through. The actual
process() function gets a NULL mask pointer in this case, and so
would composite the layer as if it had no mask, exposing areas that
should be masked out.
Add a GimpOperationLayerMode::parent_process() function, which
subclasses can override instead of GeglOperation::process(), and
make sure to update the GimpOperationLayerMode::opacity field
before calling this function (and, subsequently, before calling
GimpOperationLayerMode::process()).
Clean up the rest of the fields, and adjust the rest of the code.
(cherry picked from commit 646c804c04)
In gimp:replace, when compositing the same content over itself,
i.e., when the input and aux buffers share the same storage and
same tile alignment, pass the input buffer directly as output,
instead of doing actual processing.
In particular, this happens when processing a pass-through group
outside of its actual bounds.
(cherry picked from commit 2eaaa950a5)
In gimp:buffer-source-validate, invalidate the node in response to
GimpTileHandlerValidate::invalidated, added in the previous commit,
in addition to GeglBuffer::changed.
(cherry picked from commit bb8ee0e686)
In GimpHistogram, get rid of the "n-channels" property and
corresponding gimp_histogram_n_channels() function. The former
returned the actual number of channels, but this wasn't too useful,
as channel values may not be sequential; the latter returned the
number of components. Instead, add an "n-components" property and
a corresponding gimp_histogram_n_components() function, both of
which return the number of components. Furthermore, add a
gimp_histogram_has_channel() function, which determines if the
histogram has a given channel; this allows for simple testing for
channel availability, which was done wrong in various places.
Adjust the GimpHistogram code for the changes, and clean it up,
fixing a few bugs.
Adjust users of GimpHisotgram for the changes. In particular,
in GimpHisotgramView, fix the channel-availability test when
setting the view's histogram (which happens whenever the active
drawable's preview is frozen), to avoid erroneously swithcing the
view's channel back to "Value" when a non-RGB channel is selected.
(cherry picked from commit fc17f0ed0c)
... similarly to gimp:layer-mode, however, assume the destnation
(backdrop) is not included when the layer's opacity is 1, and
there's no mask.
(cherry picked from commit 998f89e3cb)
The latter is broken and doesn't guarantee a decimal point with the
current bug. Also, g_ascii_dtostr() doesn't need the format parameter
and produces nicer output.
(cherry picked from commit c0fb66254e)
In gimp:gradient, fix dithering to correspond to how we actually
round float values to 8-bit. In particular, this avoids
introducing noise when a component is fixed at 0 or 1 along a
segment.
(cherry picked from commit e22fcc8942)
Some blend funcs depend on constants from the specifc color space we are
operating in and needs the space or operation propagated to the worker function
of the operation as discovered in issue #3451.
This commit propagates the operation, leaving the specific blend functions
needing it to call gegl_operation_get_source_space or similar without needing
that overhead for the rest.
(cherry picked from commit 8e90468308)
Blacklist the "tools-offset" action in the GUI, and only keep
"filters-offset", to avoid duplication. Update gimp:offset's
description, so that "filters-offset" gets a proper tool-tip.
(cherry picked from commit 42d4255262)
Add a new gimp:offset operation, which implements equivalent
functionality to gimp_drawable_offset(), in preparation for adding
an interactive offset tool.
To simplify things, add a GIMP_OFFSET_WRAP_AROUND value to the
GimpOffsetType enum, to avoid the need for a separate wrap-around
flag. This makes the gimp-drawable-offset procedure parameters a
little superfluous, but whatever.
(cherry picked from commit 40fefb6076)
In gimp:mask-components, add an "alpha" property, which controls
the masked-in alpha value in case there's no aux buffer. Set it to
0 by default, so that gimp:mask-components behaves normally in the
absence of an aux buffer (as if the aux buffer was empty). Set it
to 1 in the image's visible-mask node, to maintain the current
alpha-component visibility behavior.
This fixes incorrect results when the output bounding box of a
drawable filter is smaller than the drawable, which can lead to a
NULL aux buffer being fed to the filter's gimp:mask-components
node.
(cherry picked from commit 6425bf820a)
On a second thought... nope :) We'll fix it another way.
This reverts commit 60947b7a34826657395ca485d334ccb302106e07.
(cherry picked from commit 3766af9ac9)
In GimpOperationLayerMode and GimpOperationReplace, make sure we
don't return a NULL output buffer, or forward a NULL input buffer,
but rather create an appropriate empty buffer in this case. This
avoids wrong results when the layer-mode op's output is connected
to the aux input of a subsequent op, as a result of the op behaving
differently with a NULL aux buffer (in particular, this can happen
when a drawable filter's output bounding box is smaller than the
drawable.)
(cherry picked from commit 8fcac3298c)
Extend last commit to also disregard the composite space when the
layer mode is trivial and only the source region is included in
compositing, since, in this case, the source color is unmodified.
(cherry picked from commit c2021d3c5b)
In gimp_layer_mode_get_format(), disregard the requested composite
space when selecting the format, if the input layer mode is alpha-
only, and the requested composite mode is not UNION, since, in this
case, the layer mode doesn't combine the layer/backdrop colors, and
rather only modifies the alpha of one of them. This allows us to
use the preferred format, avoiding gamma conversion.
This particularly improves the performance of the Eraser tool in
perceptual images.
(cherry picked from commit a5962e4049)
In GimpCurve, replace the use of a fixed-length control-point array
with a dynamically-sized array. Adapt GimpCurve's interface, and
the rest of the code.
In addition to simplifying the code, this fixes a bug where the
curve object could be broken by moving the mouse too fast (yep...),
and allows more accurate point placement, both in the GUI editor,
and through canvas interaction in the Curves tool (see issue #814).
(cherry picked from commit b6d829a1b2)
In gimp:fill-source, align the result rect to the drawable buffer's
tile grid, so that all tiles are COWed for solid-color fills.
(cherry picked from commit fa31854a66)
