imgIContainer.idl

/** -*- Mode: C++; tab-width: 2; indent-tabs-mode: nil; c-basic-offset: 2 -*-
 *
 * This Source Code Form is subject to the terms of the Mozilla Public
 * License, v. 2.0. If a copy of the MPL was not distributed with this
 * file, You can obtain one at http://mozilla.org/MPL/2.0/. */

#include "nsISupports.idl"

%{C++
#include "DrawResult.h"
#include "gfxContext.h"
#include "gfxMatrix.h"
#include "gfxRect.h"
#include "mozilla/gfx/2D.h"
#include "mozilla/Maybe.h"
#include "mozilla/RefPtr.h"
#include "nsRect.h"
#include "nsSize.h"
#include "limits.h"

class nsIDocument;

namespace mozilla {
namespace layers {
class LayerManager;
class ImageContainer;
}
}

class nsIFrame;

namespace mozilla {
class TimeStamp;
class SVGImageContext;
}

namespace mozilla {
namespace image {

class ImageRegion;
struct Orientation;

}
}

%}

native DrawResult(mozilla::image::DrawResult);
[ptr] native gfxContext(gfxContext);
[ref] native gfxMatrix(gfxMatrix);
[ref] native gfxRect(gfxRect);
[ref] native gfxSize(gfxSize);
native SamplingFilter(mozilla::gfx::SamplingFilter);
[ref] native nsIntRect(nsIntRect);
native nsIntRectByVal(nsIntRect);
[ref] native nsIntSize(nsIntSize);
native nsSize(nsSize);
[ptr] native nsIFrame(nsIFrame);
native TempRefImageContainer(already_AddRefed<mozilla::layers::ImageContainer>);
[ref] native ImageRegion(mozilla::image::ImageRegion);
[ptr] native LayerManager(mozilla::layers::LayerManager);
native Orientation(mozilla::image::Orientation);
[ref] native TimeStamp(mozilla::TimeStamp);
[ref] native MaybeSVGImageContext(mozilla::Maybe<mozilla::SVGImageContext>);
native TempRefSourceSurface(already_AddRefed<mozilla::gfx::SourceSurface>);
native TempRefImgIContainer(already_AddRefed<imgIContainer>);
native nsIntSizeByVal(nsIntSize);
[ptr] native nsIDocument(nsIDocument);


/**
 * imgIContainer is the interface that represents an image. It allows
 * access to frames as Thebes surfaces. It also allows drawing of images
 * onto Thebes contexts.
 *
 * Internally, imgIContainer also manages animation of images.
 */
[scriptable, builtinclass, uuid(a8dbee24-ff86-4755-b40e-51175caf31af)]
interface imgIContainer : nsISupports
{
  /**
   * The width of the container rectangle.  In the case of any error,
   * zero is returned, and an exception will be thrown.
   */
  readonly attribute int32_t width;

  /**
   * The height of the container rectangle.  In the case of any error,
   * zero is returned, and an exception will be thrown.
   */
  readonly attribute int32_t height;

  /**
   * The intrinsic size of this image in appunits. If the image has no intrinsic
   * size in a dimension, -1 will be returned for that dimension. In the case of
   * any error, an exception will be thrown.
   */
  [noscript] readonly attribute nsSize intrinsicSize;

  /**
   * The (dimensionless) intrinsic ratio of this image. In the case of any
   * error, an exception will be thrown.
   */
  [noscript] readonly attribute nsSize intrinsicRatio;

  /**
   * Given a size at which this image will be displayed, and the drawing
   * parameters affecting how it will be drawn, returns the image size which
   * should be used to draw to produce the highest quality result. This is the
   * appropriate size, for example, to use as an input to the pixel snapping
   * algorithm.
   *
   * For best results the size returned by this method should not be cached. It
   * can change over time due to changes in the internal state of the image.
   *
   * @param aDest The size of the destination rect into which this image will be
   *              drawn, in device pixels.
   * @param aWhichFrame Frame specifier of the FRAME_* variety.
   * @param aSamplingFilter The filter to be used if we're scaling the image.
   * @param aFlags Flags of the FLAG_* variety
   */
  [notxpcom, nostdcall] nsIntSizeByVal
  optimalImageSizeForDest([const] in gfxSize aDest, in uint32_t aWhichFrame,
                          in SamplingFilter aSamplingFilter, in uint32_t aFlags);

  /**
    * Enumerated values for the 'type' attribute (below).
    */
  const unsigned short TYPE_RASTER = 0;
  const unsigned short TYPE_VECTOR = 1;

  /**
   * The type of this image (one of the TYPE_* values above).
   */
  [infallible] readonly attribute unsigned short type;

  /**
   * Whether this image is animated. You can only be guaranteed that querying
   * this will not throw if STATUS_DECODE_COMPLETE is set on the imgIRequest.
   *
   * @throws NS_ERROR_NOT_AVAILABLE if the animated state cannot be determined.
   */
  readonly attribute boolean animated;

  /**
   * Flags for imgIContainer operations.
   *
   * Meanings:
   *
   * FLAG_NONE: Lack of flags.
   *
   * FLAG_SYNC_DECODE: Forces synchronous/non-progressive decode of all
   * available data before the call returns.
   *
   * FLAG_SYNC_DECODE_IF_FAST: Like FLAG_SYNC_DECODE, but requests a sync decode
   * be performed only if ImageLib estimates it can be completed very quickly.
   *
   * FLAG_ASYNC_NOTIFY: Send notifications asynchronously, even if we decode
   * synchronously beause of FLAG_SYNC_DECODE or FLAG_SYNC_DECODE_IF_FAST.
   *
   * FLAG_DECODE_NO_PREMULTIPLY_ALPHA: Do not premultiply alpha if
   * it's not already premultiplied in the image data.
   *
   * FLAG_DECODE_NO_COLORSPACE_CONVERSION: Do not do any colorspace conversion;
   * ignore any embedded profiles, and don't convert to any particular
   * destination space.
   *
   * FLAG_CLAMP: Extend the image to the fill area by clamping image sample
   * coordinates instead of by tiling. This only affects 'draw'.
   *
   * FLAG_HIGH_QUALITY_SCALING: A hint as to whether this image should be
   * scaled using the high quality scaler. Do not set this if not drawing to
   * a window or not listening to invalidations. Passing this flag will do two
   * things: 1) request a decode of the image at the size asked for by the
   * caller if one isn't already started or complete, and 2) allows a decoded
   * frame of any size (it could be neither the requested size, nor the
   * intrinsic size) to be substituted.
   *
   * FLAG_WANT_DATA_SURFACE: Can be passed to GetFrame when the caller wants a
   * DataSourceSurface instead of a hardware accelerated surface. This can be
   * important for performance (by avoiding an upload to/readback from the GPU)
   * when the caller knows they want a SourceSurface of type DATA.
   *
   * FLAG_BYPASS_SURFACE_CACHE: Forces drawing to happen rather than taking
   * cached rendering from the surface cache. This is used when we are printing,
   * for example, where we want the vector commands from VectorImages to end up
   * in the PDF output rather than a cached rendering at screen resolution.
   *
   * FLAG_FORCE_PRESERVEASPECTRATIO_NONE: Force scaling this image
   * non-uniformly if necessary. This flag is for vector image only. A raster
   * image should ignore this flag. While drawing a vector image with this
   * flag, do not force uniform scaling even if its root <svg> node has a
   * preserveAspectRatio attribute that would otherwise require uniform
   * scaling , such as xMinYMin/ xMidYMin. Always scale the graphic content of
   * the given image non-uniformly if necessary such that the image's
   * viewBox (if specified or implied by height/width attributes) exactly
   * matches the viewport rectangle.
   *
   * FLAG_FORCE_UNIFORM_SCALING: Signal to ClippedImage::OptimalSizeForDest that
   * its returned size can only scale the image's size *uniformly* (by the same
   * factor in each dimension). We need this flag when painting border-image
   * section with SVG image source-data, if the SVG image has no viewBox and no
   * intrinsic size. In such a case, we synthesize a viewport for the SVG image
   * (a "window into SVG space") based on the border image area, and we need to
   * be sure we don't subsequently scale that viewport in a way that distorts
   * its contents by stretching them more in one dimension than the other.
   */
  const unsigned long FLAG_NONE                            = 0x0;
  const unsigned long FLAG_SYNC_DECODE                     = 0x1;
  const unsigned long FLAG_SYNC_DECODE_IF_FAST             = 0x2;
  const unsigned long FLAG_ASYNC_NOTIFY                    = 0x4;
  const unsigned long FLAG_DECODE_NO_PREMULTIPLY_ALPHA     = 0x8;
  const unsigned long FLAG_DECODE_NO_COLORSPACE_CONVERSION = 0x10;
  const unsigned long FLAG_CLAMP                           = 0x20;
  const unsigned long FLAG_HIGH_QUALITY_SCALING            = 0x40;
  const unsigned long FLAG_WANT_DATA_SURFACE               = 0x80;
  const unsigned long FLAG_BYPASS_SURFACE_CACHE            = 0x100;
  const unsigned long FLAG_FORCE_PRESERVEASPECTRATIO_NONE  = 0x200;
  const unsigned long FLAG_FORCE_UNIFORM_SCALING           = 0x400;

  /**
   * A constant specifying the default set of decode flags (i.e., the default
   * values for FLAG_DECODE_*).
   */
  const unsigned long DECODE_FLAGS_DEFAULT = 0;

  /**
    * Constants for specifying various "special" frames.
    *
    * FRAME_FIRST: The first frame
    * FRAME_CURRENT: The current frame
    *
    * FRAME_MAX_VALUE should be set to the value of the maximum constant above,
    * as it is used for ensuring that a valid value was passed in.
    */
  const unsigned long FRAME_FIRST = 0;
  const unsigned long FRAME_CURRENT = 1;
  const unsigned long FRAME_MAX_VALUE = 1;

  /**
   * Get a surface for the given frame. This may be a platform-native,
   * optimized surface, so you cannot inspect its pixel data. If you
   * need that, use SourceSurface::GetDataSurface.
   *
   * @param aWhichFrame Frame specifier of the FRAME_* variety.
   * @param aFlags Flags of the FLAG_* variety
   */
  [noscript, notxpcom] TempRefSourceSurface getFrame(in uint32_t aWhichFrame,
                                                     in uint32_t aFlags);

  /**
   * Get a surface for the given frame at the specified size. Matching the
   * requested size is best effort; it's not guaranteed that the surface you get
   * will be a perfect match. (Some reasons you may get a surface of a different
   * size include: if you requested upscaling, if downscale-during-decode is
   * disabled, or if you didn't request the first frame.)
   *
   * @param aSize The desired size.
   * @param aWhichFrame Frame specifier of the FRAME_* variety.
   * @param aFlags Flags of the FLAG_* variety
   */
  [noscript, notxpcom] TempRefSourceSurface getFrameAtSize([const] in nsIntSize aSize,
                                                           in uint32_t aWhichFrame,
                                                           in uint32_t aFlags);

  /**
   * Returns true if this image will draw opaquely right now if asked to draw
   * with FLAG_HIGH_QUALITY_SCALING and otherwise default flags. If this image
   * (when decoded) is opaque but no decoded frames are available then
   * willDrawOpaqueNow will return false.
   */
  [noscript, notxpcom] boolean willDrawOpaqueNow();

  /**
   * @return true if getImageContainer() is expected to return a valid
   *         ImageContainer when passed the given @Manager and @Flags
   *         parameters.
   */
  [noscript, notxpcom] boolean isImageContainerAvailable(in LayerManager aManager,
                                                         in uint32_t aFlags);
  /**
   * Attempts to create an ImageContainer (and Image) containing the current
   * frame.
   *
   * Avoid calling this unless you're actually going to layerize this image.
   *
   * @param aManager The LayerManager which will be used to create the
   *                 ImageContainer.
   * @param aFlags Decoding / drawing flags (in other words, FLAG_* flags).
   *               Currently only FLAG_SYNC_DECODE and FLAG_SYNC_DECODE_IF_FAST
   *               are supported.
   * @return An ImageContainer for the current frame, or nullptr if one could
   *         not be created.
   */
  [noscript, notxpcom] TempRefImageContainer getImageContainer(in LayerManager aManager,
                                                               in uint32_t aFlags);

  /**
   * Draw the requested frame of this image onto the context specified.
   *
   * Drawing an image involves scaling it to a certain size (which may be
   * implemented as a "smart" scale by substituting an HQ-scaled frame or
   * rendering at a high DPI), and then selecting a region of that image to
   * draw. That region is drawn onto the graphics context and in the process
   * transformed by the context matrix, which determines the final area that is
   * filled. The basic process looks like this:
   *
   *                           +------------------+
   *                           |      Image       |
   *                           |                  |
   *                           | intrinsic width  |
   *                           |        X         |
   *                           | intrinsic height |
   *                           +------------------+
   *                          /                    \
   *                         /                      \
   *                        /    (scale to aSize)    \
   *                       /                          \
   *                      +----------------------------+
   *                      |                            |
   *                      |        Scaled Image        |
   *                      | aSize.width X aSize.height |
   *                      |                            |
   *                      |       +---------+          |
   *                      |       | aRegion |          |
   *                      |       +---------+          |
   *                      +-------(---------(----------+
   *                              |         |
   *                             /           \
   *                            |  (transform |
   *                           /  by aContext  \
   *                          |     matrix)     |
   *                         /                   \
   *                        +---------------------+
   *                        |                     |
   *                        |      Fill Rect      |
   *                        |                     |
   *                        +---------------------+
   *
   * The region may extend outside of the scaled image's boundaries. It's
   * actually a region in tiled image space, which is formed by tiling the
   * scaled image infinitely in every direction. Drawing with a region larger
   * than the scaled image thus causes the filled area to contain multiple tiled
   * copies of the image, which looks like this:
   *
   *           ....................................................
   *           :                :                :                :
   *           :      Tile      :      Tile      :      Tile      :
   *           :        +------------[aRegion]------------+       :
   *           :........|.......:................:........|.......:
   *           :        |       :                :        |       :
   *           :      Ti|le     :  Scaled Image  :      Ti|le     :
   *           :        |       :                :        |       :
   *           :........|.......:................:........|.......:
   *           :        +---------------------------------+       :
   *           :      Ti|le     :      Tile      :      Ti|le     :
   *           :       /        :                :         \      :
   *           :......(.........:................:..........).....:
   *                  |                                     |
   *                 /                                       \
   *                |      (transform by aContext matrix)     |
   *               /                                           \
   *              +---------------------------------------------+
   *              |     :                                 :     |
   *              |.....:.................................:.....|
   *              |     :                                 :     |
   *              |     :           Tiled Fill            :     |
   *              |     :                                 :     |
   *              |.....:.................................:.....|
   *              |     :                                 :     |
   *              +---------------------------------------------+
   *
   *
   * @param aContext The Thebes context to draw the image to.
   * @param aSize The size to which the image should be scaled before drawing.
   *              This requirement may be satisfied using HQ scaled frames,
   *              selecting from different resolution layers, drawing at a
   *              higher DPI, or just performing additional scaling on the
   *              graphics context. Callers can use optimalImageSizeForDest()
   *              to determine the best choice for this parameter if they have
   *              no special size requirements.
   * @param aRegion The region in tiled image space which will be drawn onto the
   *                graphics context. aRegion is in the coordinate space of the
   *                image after it has been scaled to aSize - that is, the image
   *                is scaled first, and then aRegion is applied. When aFlags
   *                includes FLAG_CLAMP, the image will be extended to this area
   *                by clamping image sample coordinates. Otherwise, the image
   *                will be automatically tiled as necessary. aRegion can also
   *                optionally contain a second region which restricts the set
   *                of pixels we're allowed to sample from when drawing; this
   *                is only of use to callers which need to draw with pixel
   *                snapping.
   * @param aWhichFrame Frame specifier of the FRAME_* variety.
   * @param aSamplingFilter The filter to be used if we're scaling the image.
   * @param aSVGContext If specified, SVG-related rendering context, such as
   *                    overridden attributes on the image document's root <svg>
   *                    node, and the size of the viewport that the full image
   *                    would occupy. Ignored for raster images.
   * @param aFlags Flags of the FLAG_* variety
   * @return A DrawResult value indicating whether and to what degree the
   *         drawing operation was successful.
   */
  [noscript, notxpcom] DrawResult
  draw(in gfxContext aContext,
       [const] in nsIntSize aSize,
       [const] in ImageRegion aRegion,
       in uint32_t aWhichFrame,
       in SamplingFilter aSamplingFilter,
       [const] in MaybeSVGImageContext aSVGContext,
       in uint32_t aFlags,
       in float aOpacity);

  /*
   * Ensures that an image is decoding. Calling this function guarantees that
   * the image will at some point fire off decode notifications. Images that
   * can be decoded "quickly" according to some heuristic will be decoded
   * synchronously.
   *
   * @param aFlags Flags of the FLAG_* variety. Only FLAG_ASYNC_NOTIFY
   *               is accepted; all others are ignored.
   */
  [noscript] void startDecoding(in uint32_t aFlags);

  /*
   * Exactly like startDecoding above except returns whether the current frame
   * of the image is complete or not.
   *
   * @param aFlags Flags of the FLAG_* variety. Only FLAG_ASYNC_NOTIFY
   *               is accepted; all others are ignored.
   */
  [noscript, notxpcom] boolean startDecodingWithResult(in uint32_t aFlags);

  /*
   * This method triggers decoding for an image, but unlike startDecoding() it
   * enables the caller to provide more detailed information about the decode
   * request.
   *
   * @param aSize The size to which the image should be scaled while decoding,
   *              if possible. If the image cannot be scaled to this size while
   *              being decoded, it will be decoded at its intrinsic size.
   * @param aFlags Flags of the FLAG_* variety. Only the decode flags
   *               (FLAG_DECODE_*) and FLAG_SYNC_DECODE (which will
   *               synchronously decode images that can be decoded "quickly",
   *               just like startDecoding() does) are accepted; others will be
   *               ignored.
   */
  [noscript] void requestDecodeForSize([const] in nsIntSize aSize,
                                       in uint32_t aFlags);

  /**
    * Increments the lock count on the image. An image will not be discarded
    * as long as the lock count is nonzero. Note that it is still possible for
    * the image to be undecoded if decode-on-draw is enabled and the image
    * was never drawn.
    *
    * Upon instantiation images have a lock count of zero.
    */
  void lockImage();

  /**
    * Decreases the lock count on the image. If the lock count drops to zero,
    * the image is allowed to discard its frame data to save memory.
    *
    * Upon instantiation images have a lock count of zero. It is an error to
    * call this method without first having made a matching lockImage() call.
    * In other words, the lock count is not allowed to be negative.
    */
  void unlockImage();

  /**
   * If this image is unlocked, discard its decoded data.  If the image is
   * locked or has already been discarded, do nothing.
   */
  void requestDiscard();

  /**
    * Indicates that this imgIContainer has been triggered to update
    * its internal animation state. Likely this should only be called
    * from within nsImageFrame or objects of similar type.
    */
  [notxpcom] void requestRefresh([const] in TimeStamp aTime);

  /**
   * Animation mode Constants
   *   0 = normal
   *   1 = don't animate
   *   2 = loop once
   */
  const short kNormalAnimMode   = 0;
  const short kDontAnimMode     = 1;
  const short kLoopOnceAnimMode = 2;

  attribute unsigned short animationMode;

  /* Methods to control animation */
  void resetAnimation();

  /*
   * Returns an index for the requested animation frame (either FRAME_FIRST or
   * FRAME_CURRENT).
   *
   * The units of the index aren't specified, and may vary between different
   * types of images. What you can rely on is that on all occasions when
   * getFrameIndex(FRAME_CURRENT) returns a certain value,
   * draw(..FRAME_CURRENT..) will draw the same frame. The same holds for
   * FRAME_FIRST as well.
   *
   * @param aWhichFrame Frame specifier of the FRAME_* variety.
   */
  [notxpcom] float getFrameIndex(in uint32_t aWhichFrame);

  /*
   * Returns the inherent orientation of the image, as described in the image's
   * metadata (e.g. EXIF).
   */
  [notxpcom] Orientation getOrientation();

  /*
   * Returns the delay, in ms, between the first and second frame. If this
   * returns 0, there is no delay between first and second frame (i.e., this
   * image could render differently whenever it draws).
   *
   * If this image is not animated, or not known to be animated (see attribute
   * animated), returns -1.
   */
  [notxpcom] int32_t getFirstFrameDelay();

  /*
   * If this is an animated image that hasn't started animating already, this
   * sets the animation's start time to the indicated time.
   *
   * This has no effect if the image isn't animated or it has started animating
   * already; it also has no effect if the image format doesn't care about
   * animation start time.
   *
   * In all cases, animation does not actually begin until startAnimation(),
   * resetAnimation(), or requestRefresh() is called for the first time.
   */
  [notxpcom] void setAnimationStartTime([const] in TimeStamp aTime);

  /*
   * Given an invalidation rect in the coordinate system used by the decoder,
   * returns an invalidation rect in image space.
   *
   * This is the identity transformation in most cases, but the result can
   * differ if the image is wrapped by an ImageWrapper that changes its size
   * or orientation.
   */
  [notxpcom] nsIntRectByVal
  getImageSpaceInvalidationRect([const] in nsIntRect aRect);

  /*
   * Removes any ImageWrappers and returns the unwrapped base image.
   */
  [notxpcom, nostdcall] TempRefImgIContainer unwrap();

  /*
   * Propagate the use counters (if any) from this container to the passed in
   * document.
   */
  [noscript, notxpcom] void propagateUseCounters(in nsIDocument aDocument);

  %{C++
  /*
   * Get the set of sizes the image can decode to natively.
   */
  virtual nsresult GetNativeSizes(nsTArray<nsIntSize>& aNativeSizes) const = 0;
  %}
};