HTML Canvas 2D Context, Level 2

A CanvasRenderingContext2D object can be obtained in two ways: the getContext() method on a canvas element (which invokes the 2D context creation algorithm), and the CanvasRenderingContext2D() constructor.

A CanvasRenderingContext2D object has a scratch bitmap and can be bound to an output bitmap. These are initialized when the object is created, and can be subsequently adjusted when the rendering context is bound or unbound. In some cases, these bitmaps are the same underlying bitmap. In general, the scratch bitmap is what scripts interact with, and the output bitmap is what is being displayed. These bitmaps always have the same dimensions.

Each such bitmap has an origin-clean flag, which can be set to true or false. Initially, when one of these bitmaps is created, its origin-clean flag must be set to true.

These bitmaps also have a hit region list, which is described in a later section. Initially, this list is empty. Scratch bitmaps also have a list of pending interface actions, which can contain instructions to draw the user's attention to a location on the bitmap, and instructions to scroll to a location on the bitmap. Initially, this list is also empty.

The CanvasRenderingContext2D 2D rendering context represents a flat linear Cartesian surface whose origin (0,0) is at the top left corner, with the coordinate space having x values increasing when going right, and y values increasing when going down. The x-coordinate of the right-most edge is equal to the width of the rendering context's scratch bitmap in CSS pixels; similarly, the y-coordinate of the bottom-most edge is equal to the height of the rendering context's scratch bitmap in CSS pixels.

The size of the coordinate space does not necessarily represent the size of the actual bitmaps that the user agent will use internally or during rendering. On high-definition displays, for instance, the user agent may internally use bitmaps with two device pixels per unit in the coordinate space, so that the rendering remains at high quality throughout. Anti-aliasing can similarly be implemented using over-sampling with bitmaps of a higher resolution than the final image on the display.

The 2D context creation algorithm, which is passed a target (a canvas element), consists of running the following steps:

1. Create a new CanvasRenderingContext2D object.

2. Initialize its canvas attribute to point to target.

3. Let the new CanvasRenderingContext2D object's output bitmap and scratch bitmap both be the same bitmap as target's bitmap (so that they are shared).

4. Set bitmap dimensions to the numeric values of target's width and height content attributes.

5. Return the new CanvasRenderingContext2D object.

The CanvasRenderingContext2D() constructor, when invoked, must run the following steps:

1. Create a new CanvasRenderingContext2D object.

2. Initialize its canvas attribute to null.

3. Let the new CanvasRenderingContext2D object's scratch bitmap be a new bitmap.

4. If the constructor was called with arguments, let width and height be the first and second arguments, respectively. Otherwise, let width and height be 300 and 150, respectively.

5. Set bitmap dimensions to width and height.

6. Let the new CanvasRenderingContext2D object have no output bitmap.

7. Return the new CanvasRenderingContext2D object.

When the user agent is required to commit the scratch bitmap for a rendering context, it must run the following steps:

1. Let bitmap copy be a copy of the rendering context's scratch bitmap.

2. Let origin-clean flag copy be a copy of the rendering context's scratch bitmap's origin-clean flag.

3. Let hit region list copy be a copy of the rendering context's scratch bitmap's hit region list.

4. Let list of pending interface actions copy be a copy of the rendering context's scratch bitmap's list of pending interface actions.

5. Empty the scratch bitmap's list of pending interface actions.

6. If the rendering context has no output bitmap, abort these steps.

7. Let output bitmap be the rendering context's output bitmap.

8. Let canvas be the canvas element to which the rendering context was most recently bound.

9. Queue a task associated with canvas' Document to perform the following substeps:

   1. Overwrite output bitmap with bitmap copy.

   2. Overwrite output bitmap's origin-clean flag with origin-clean flag copy.

   3. Overwrite output bitmap's hit region list with hit region list copy.

   4. Follow the directions in the list of pending interface actions copy.

The algorithm above must use the canvas updating task source (which is just used by this algorithm).

The commit() method must run the following steps:

1. If the rendering context's context bitmap mode is fixed, throw an InvalidStateError exception and abort these steps.

2. Commit the scratch bitmap for the rendering context.

The scratch bitmap is only committed when the commit() method is called. (This doesn't matter for canvas elements in direct-2d mode, since there the scratch bitmap is also the canvas element's bitmap so every drawing operation is immediately drawn.)

When the user agent is to set bitmap dimensions to width and height, it must run the following steps:

1. Reset the rendering context to its default state.

2. Clear the scratch bitmap's hit region list and its list of pending interface actions.

3. Resize the scratch bitmap to the new width and height and clear it to fully transparent black.

4. If the rendering context has an output bitmap, and the scratch bitmap is a different bitmap than the output bitmap, then resize the output bitmap to the new width and height and clear it to fully transparent black.

5. If the rendering context's context bitmap mode is fixed, then run these substeps:

   1. Let canvas be the canvas element to which the rendering context's canvas attribute was initialized.

   2. If the rendering context's context bitmap mode is fixed and the numeric value of the canvas' width content attribute differs from width, then set canvas' width content attribute to the shortest possible string representing width as a valid non-negative integer.

   3. If the rendering context's context bitmap mode is fixed and the numeric value of the canvas' height content attribute differs from height, then set canvas' height content attribute to the shortest possible string representing height as a valid non-negative integer.

// canvas is a reference to a <canvas> element
var context = canvas.getContext('2d');
context.fillRect(0,0,50,50);
canvas.setAttribute('width', '300'); // clears the canvas
context.fillRect(0,100,50,50);
canvas.width = canvas.width; // clears the canvas
context.fillRect(100,0,50,50); // only this square remains

When the user agent is to run the unbinding steps for a rendering context, it must run the following steps:

1. Reset the rendering context to its default state.

2. Clear the scratch bitmap's hit region list and its list of pending interface actions.

3. Clear the CanvasRenderingContext2D object's scratch bitmap to a transparent black.

4. Set the CanvasRenderingContext2D object's scratch bitmap's origin-clean flag to true.

5. Let the CanvasRenderingContext2D object have no output bitmap.

When the user agent is to run the binding steps to bind the rendering context to the canvas element target, it must run the following steps:

1. Reset the rendering context to its default state.

2. Clear the scratch bitmap's hit region list and its list of pending interface actions.

3. Resize the CanvasRenderingContext2D object's scratch bitmap to the dimensions of target's bitmap and clear it to fully transparent black.

4. Set the CanvasRenderingContext2D object's scratch bitmap's origin-clean flag to true.

5. Let the CanvasRenderingContext2D object's output bitmap be target's bitmap.

The canvas attribute must return the value it was initialized to when the object was created.

The width attribute, on getting, must return the width of the rendering context's scratch bitmap, in CSS pixels. On setting, it must set bitmap dimensions to the new value and the current height of the rendering context's scratch bitmap in CSS pixels, respectively.

The height attribute, on getting, must return the height of the rendering context's scratch bitmap, in CSS pixels. On setting, it must set bitmap dimensions to the current width of the rendering context's scratch bitmap in CSS pixels and the new value, respectively.

Except where otherwise specified, for the 2D context interface, any method call with a numeric argument whose value is infinite or a NaN value must be ignored.

Whenever the CSS value currentColor is used as a color in the CanvasRenderingContext2D API, the "computed value of the 'color' property" for the purposes of determining the computed value of the currentColor keyword is the value described by the appropriate entry in the following list:

If the rendering context's context bitmap mode is fixed and the canvas element is being rendered

The "computed value of the 'color' property" for the purposes of determining the computed value of the currentColor keyword is the computed value of the 'color' property on the canvas element at the time that the color is specified (e.g. when the appropriate attribute is set, or when the method is called; not when the color is rendered or otherwise used). [CSSCOLOR]

In all other cases

The "computed value of the 'color' property" for the purposes of determining the computed value of the currentColor keyword is fully opaque black. [CSSCOLOR]

In the case of addColorStop() on CanvasGradient, the "computed value of the 'color' property" for the purposes of determining the computed value of the currentColor keyword is always fully opaque black (there is no associated element). [CSSCOLOR]

This is because CanvasGradient objects are canvas-neutral — a CanvasGradient object created by one canvas can be used by another, and there is therefore no way to know which is the "element in question" at the time that the color is specified.

Similar concerns exist with font-related properties; the rules for those are described in detail in the relevant section below.

The save() method must push a copy of the current drawing state onto the drawing state stack.

The restore() method must pop the top entry in the drawing state stack, and reset the drawing state it describes. If there is no saved state, the method must do nothing.

When the user agent is to reset the rendering context to its default state, it must clear the drawing state stack and everything that drawing state consists of to initial values.

Each DrawingStyle object can have a styles scope object.

The DrawingStyle() constructor, when invoked, must return a newly created DrawingStyle object. If the constructor was passed an argument, then the DrawingStyle object's styles scope object is that element. Otherwise, if the JavaScript global environment is a document environment, the object's styles scope object is the Document object of the active document of the browsing context of the Window object on which the interface object of the invoked constructor is found. Otherwise, the JavaScript global environment is a worker environment, and the styles scope object is the worker.

Objects that implement the CanvasDrawingStyles interface have attributes and methods (defined in this section) that control how lines are treated by the object.

The lineWidth attribute gives the width of lines, in coordinate space units. On getting, it must return the current value. On setting, zero, negative, infinite, and NaN values must be ignored, leaving the value unchanged; other values must change the current value to the new value.

When the object implementing the CanvasDrawingStyles interface is created, the lineWidth attribute must initially have the value 1.0.

The lineCap attribute defines the type of endings that UAs will place on the end of lines. The three valid values are butt, round, and square.

On getting, it must return the current value. On setting, if the new value is one of the literal strings butt, round, and square, then the current value must be changed to the new value; other values must ignored, leaving the value unchanged.

When the object implementing the CanvasDrawingStyles interface is created, the lineCap attribute must initially have the value butt.

The lineJoin attribute defines the type of corners that UAs will place where two lines meet. The three valid values are bevel, round, and miter.

On getting, it must return the current value. On setting, if the new value is one of the literal strings bevel, round, and miter, then the current value must be changed to the new value; other values must be ignored, leaving the value unchanged.

When the object implementing the CanvasDrawingStyles interface is created, the lineJoin attribute must initially have the value miter.

When the lineJoin attribute has the value miter, strokes use the miter limit ratio to decide how to render joins. The miter limit ratio can be explicitly set using the miterLimit attribute. On getting, it must return the current value. On setting, zero, negative, infinite, and NaN values must be ignored, leaving the value unchanged; other values must change the current value to the new value.

When the object implementing the CanvasDrawingStyles interface is created, the miterLimit attribute must initially have the value 10.0.

Each CanvasDrawingStyles object has a dash list, which is either empty or consists of an even number of non-negative numbers. Initially, the dash list must be empty.

When the setLineDash() method is invoked, it must run the following steps:

1. Let a be the argument.

2. If any value in a is not finite (e.g. an Infinity or a NaN value), or if any value is negative (less than zero), then abort these steps (without throwing an exception; user agents could show a message on a developer console, though, as that would be helpful for debugging).

3. If the number of elements in a is odd, then let a be the concatentation of two copies of a.

4. Let the object's dash list be a.

When the getLineDash() method is invoked, it must return a sequence whose values are the values of the object's dash list, in the same order.

It is sometimes useful to change the "phase" of the dash pattern, e.g. to achieve a "marching ants" effect. The phase can be set using the lineDashOffset attribute. On getting, it must return the current value. On setting, infinite and NaN values must be ignored, leaving the value unchanged; other values must change the current value to the new value.

When the object implementing the CanvasDrawingStyles interface is created, the lineDashOffset attribute must initially have the value 0.0.

<!DOCTYPE html>
 <html>
<head>
  <title>Dash lines</title>

</head>
<body>
  <canvas id="newCanvas" width="400" height="350">This browser or document mode doesn't support canvas</canvas> 
    <script type="text/javascript">
      var Offset = 0;  // Starting value for ant offset
      var dashList = [12, 3, 3, 3];  // Create a dot/dash sequence

      function draw() {
        var canvas = document.getElementById("newCanvas");
        if (canvas.getContext) {
          var context = canvas.getContext("2d");
          context.clearRect(0, 0, canvas.width, canvas.height);
          //  set dashList for the dash sequence
          context.setLineDash(dashList);
          //  Set the current offset 
          context.lineDashOffset = Offset;
          context.lineJoin = "round";
          context.lineWidth = "3";
          context.strokeStyle = "green";
          context.strokeRect(5, 5, 300, 250);  
        }
      }

      function dodashes() {
        Offset++;
        if (Offset > 20)  // Reset offset after total of dashList values
        {
          antOffset = 0;
        }
        draw();
        setTimeout(dodashes, 30);  //  Set a leisurely march of 30ms               
      }
      dodashes();  //  Start the demo 
  </script>
</body>
</html>

When a user agent is to trace a path, given an object style that implements the CanvasDrawingStyles interface, it must run the following algorithm. This algorithm returns a new path.

1. Let path be a copy of the path being traced.

2. Remove from path any subpaths containing no lines (i.e. subpaths with just one point).

3. Replace each point in each subpath of path other than the first point and the last point of each subpath by a join that joins the line leading to that point to the line leading out of that point, such that the subpaths all consist of two points (a starting point with a line leading out of it, and an ending point with a line leading into it), one or more lines (connecting the points and the joins), and zero or more joins (each connecting one line to another), connected together such that each subpath is a series of one or more lines with a join between each one and a point on each end.

4. Add a straight closing line to each closed subpath in path connecting the last point and the first point of that subpath; change the last point to a join (from the previously last line to the newly added closing line), and change the first point to a join (from the newly added closing line to the first line).

5. If the styles dash list is empty, jump to the step labeled convert.

6. Let pattern width be the concatenation of all the entries of the styles dash list, in coordinate space units.

7. For each subpath subpath in path, run the following substeps. These substeps mutate the subpaths in path in vivo.

   1. Let subpath width be the length of all the lines of subpath, in coordinate space units.

   2. Let offset be the value of the styles lineDashOffset, in coordinate space units.

   3. While offset is greater than pattern width, decrement it by pattern width.

      While offset is less than pattern width, increment it by pattern width.

   4. Define L to be a linear coordinate line defined along all lines in subpath, such that the start of the first line in the subpath is defined as coordinate 0, and the end of the last line in the subpath is defined as coordinate width.

   5. Let position be zero minus offset.

   6. Let index be 0.

   7. Let current state be off (the other states being on and zero-on).

   8. Dash on: Let segment length be the value of the styles dash list's indexth entry.

   9. Increment position by segment length.

   10. If position is greater than subpath width, then end these substeps for this subpath and start them again for the next subpath; if there are no more subpaths, then jump to the step labeled convert instead.

   11. If segment length is non-zero, let current state be on.

   12. Increment index by one.

   13. Dash off: Let segment length be the value of the styles dash list's indexth entry.

   14. Let start be the offset position on L.

   15. Increment position by segment length.

   16. If position is less than zero, then jump to the step labeled post-cut.

   17. If start is less than zero, then let start be zero.

   18. If position is greater than subpath width, then let end be the offset subpath width on L. Otherwise, let end be the offset position on L.

   19. Jump to the first appropriate step:

       If segment length is zero and current state is off

       Do nothing, just continue to the next step.

       If current state is off

       Cut the line on which end finds itself short at end and place a point there, cutting the subpath that it was in in two; remove all line segments, joins, points, and subpaths that are between start and end; and finally place a single point at start with no lines connecting to it.

       The point has a directionality for the purposes of drawing line caps (see below). The directionality is the direction that the original line had at that point (i.e. when L was defined above).

       Otherwise

       Cut the line on which start finds itself into two at start and place a point there, cutting the subpath that it was in in two, and similarly cut the line on which end finds itself short at end and place a point there, cutting the subpath that it was in in two, and then remove all line segments, joins, points, and subpaths that are between start and end.

       If start and end are the same point, then this results in just the line being cut in two and two points being inserted there, with nothing being removed, unless a join also happens to be at that point, in which case the join must be removed.

   20. Post-cut: If position is greater than width, then jump to the step labeled convert.

   21. If segment length is greater than zero, let positioned-at-on-dash be false.

   22. Increment index by one. If it is equal to the number of entries in the styles dash list, then let index be 0.

   23. Return to the step labeled dash on.

8. Convert: This is the step that converts the path to a new path that represents its stroke.

   Create a new path that describes the edge of the areas that would be covered if a straight line of length equal to the styles lineWidth was swept along each path in path while being kept at an angle such that the line is orthogonal to the path being swept, replacing each point with the end cap necessary to satisfy the styles lineCap attribute as described previously and elaborated below, and replacing each join with the join necessary to satisfy the styles lineJoin type, as defined below.

   Caps: Each point has a flat edge perpendicular to the direction of the line coming out of it. This is them augmented according to the value of the styles lineCap. The butt value means that no additional line cap is added. The round value means that a semi-circle with the diameter equal to the styles lineWidth width must additionally be placed on to the line coming out of each point. The square value means that a rectangle with the length of the styles lineWidth width and the width of half the styles lineWidth width, placed flat against the edge perpendicular to the direction of the line coming out of the point, must be added at each point.

   Points with no lines coming out of them must have two caps placed back-to-back as if it was really two points connected to each other by an infinitesimally short straight line in the direction of the point's directionality (as defined above).

   Joins: In addition to the point where a join occurs, two additional points are relevant to each join, one for each line: the two corners found half the line width away from the join point, one perpendicular to each line, each on the side furthest from the other line.

   A filled triangle connecting these two opposite corners with a straight line, with the third point of the triangle being the join point, must be added at all joins. The lineJoin attribute controls whether anything else is rendered. The three aforementioned values have the following meanings:

   The bevel value means that this is all that is rendered at joins.

   The round value means that a filled arc connecting the two aforementioned corners of the join, abutting (and not overlapping) the aforementioned triangle, with the diameter equal to the line width and the origin at the point of the join, must be added at joins.

   The miter value means that a second filled triangle must (if it can given the miter length) be added at the join, with one line being the line between the two aforementioned corners, abutting the first triangle, and the other two being continuations of the outside edges of the two joining lines, as long as required to intersect without going over the miter length.

   The miter length is the distance from the point where the join occurs to the intersection of the line edges on the outside of the join. The miter limit ratio is the maximum allowed ratio of the miter length to half the line width. If the miter length would cause the miter limit ratio (as set by the style miterLimit attribute) to be exceeded, this second triangle must not be added.

   Subpaths in the newly created path must wind clockwise, regardless of the direction of paths in path.

9. Return the newly created path.

Objects that implement the CanvasDrawingStyles interface have attributes (defined in this section) that control how text is laid out (rasterized or outlined) by the object. Such objects can also have a font style source object. For CanvasRenderingContext2D objects whose context bitmap mode is fixed, this is their canvas element; for other CanvasRenderingContext2D objects, if the JavaScript global environment is a document environment, the object's font style source object is the Document object of the active document of the browsing context of the Window object on which the interface object of the CanvasRenderingContext2D object is found; otherwise the JavaScript global environment is a worker environment and the font style source object is the worker. For DrawingStyle objects, it's the styles scope object.

The font IDL attribute, on setting, must be parsed the same way as the 'font' property of CSS (but without supporting property-independent style sheet syntax like 'inherit'), and the resulting font must be assigned to the context, with the 'line-height' component forced to 'normal', with the 'font-size' component converted to CSS pixels, and with system fonts being computed to explicit values. If the new value is syntactically incorrect (including using property-independent style sheet syntax like 'inherit' or 'initial'), then it must be ignored, without assigning a new font value. [CSS]

Font family names must be interpreted in the context of the font style source object when the font is to be used; any fonts embedded using @font-face or loaded using the FontLoader that are visible to the font style source object must therefore be available once they are loaded. If a font is used before it is fully loaded, or if the font style source object does not have that font in scope at the time the font is to be used, then it must be treated as if it was an unknown font, falling back to another as described by the relevant CSS specifications. [CSSFONTS] [CSSFONTLOAD]

On getting, the font attribute must return the serialized form of the current font of the context (with no 'line-height' component). [CSSOM]

context.font = 'italic 400 12px/2 Unknown Font, sans-serif';

When the object implementing the CanvasDrawingStyles interface is created, the font of the context must be set to 10px sans-serif. When the 'font-size' component is set to lengths using percentages, 'em' or 'ex' units, or the 'larger' or 'smaller' keywords, these must be interpreted relative to the computed value of the 'font-size' property of the font style source object at the time that the attribute is set, if it is an element. When the 'font-weight' component is set to the relative values 'bolder' and 'lighter', these must be interpreted relative to the computed value of the 'font-weight' property of the font style source object at the time that the attribute is set, if it is an element. If the computed values are undefined for a particular case (e.g. because the font style source object is not an element or is not being rendered), then the relative keywords must be interpreted relative to the normal-weight 10px sans-serif default.

The textAlign IDL attribute, on getting, must return the current value. On setting, if the value is one of start, end, left, right, or center, then the value must be changed to the new value. Otherwise, the new value must be ignored. When the object implementing the CanvasDrawingStyles interface is created, the textAlign attribute must initially have the value start.

The textBaseline IDL attribute, on getting, must return the current value. On setting, if the value is one of top, hanging, middle, alphabetic, ideographic, or bottom, then the value must be changed to the new value. Otherwise, the new value must be ignored. When the object implementing the CanvasDrawingStyles interface is created, the textBaseline attribute must initially have the value alphabetic.

The direction IDL attribute, on getting, must return the current value. On setting, if the value is one of ltr, rtl, or inherit, then the value must be changed to the new value. Otherwise, the new value must be ignored. When the object implementing the CanvasDrawingStyles interface is created, the direction attribute must initially have the value inherit.

The following methods allow authors to manipulate the paths of objects implementing the CanvasPathMethods interface.

For CanvasRenderingContext2D objects, the points passed to the methods, and the resulting lines added to current default path by these methods, must be transformed according to the current transformation matrix before being added to the path.

The moveTo(x, y) method must create a new subpath with the specified point as its first (and only) point.

When the user agent is to ensure there is a subpath for a coordinate (x, y) on a path, the user agent must check to see if the path has any subpaths, and if it does not, then the user agent must create a new subpath with the point (x, y) as its first (and only) point, as if the moveTo() method had been called.

The closePath() method must do nothing if the object's path has no subpaths. Otherwise, it must mark the last subpath as closed, create a new subpath whose first point is the same as the previous subpath's first point, and finally add this new subpath to the path.

If the last subpath had more than one point in its list of points, then this is equivalent to adding a straight line connecting the last point back to the first point, thus "closing" the shape, and then repeating the last (possibly implied) moveTo() call.

New points and the lines connecting them are added to subpaths using the methods described below. In all cases, the methods only modify the last subpath in the object's path.

The lineTo(x, y) method must ensure there is a subpath for (x, y) if the object's path has no subpaths. Otherwise, it must connect the last point in the subpath to the given point (x, y) using a straight line, and must then add the given point (x, y) to the subpath.

The quadraticCurveTo(cpx, cpy, x, y) method must ensure there is a subpath for (cpx, cpy), and then must connect the last point in the subpath to the given point (x, y) using a quadratic Bézier curve with control point (cpx, cpy), and must then add the given point (x, y) to the subpath. [BEZIER]

The bezierCurveTo(cp1x, cp1y, cp2x, cp2y, x, y) method must ensure there is a subpath for (cp1x, cp1y), and then must connect the last point in the subpath to the given point (x, y) using a cubic Bézier curve with control points (cp1x, cp1y) and (cp2x, cp2y). Then, it must add the point (x, y) to the subpath. [BEZIER]

The arcTo(x1, y1, x2, y2, radiusX, radiusY, rotation) method must first ensure there is a subpath for (x1, y1). Then, the behavior depends on the arguments and the last point in the subpath, as described below.

Negative values for radiusX or radiusY must cause the implementation to throw an IndexSizeError exception. If radiusY is omitted, user agents must act as if it had the same value as radiusX.

Let the point (x0, y0) be the last point in the subpath, transformed by the inverse of the current transformation matrix (so that it is in the same coordinate system as the points passed to the method).

If the point (x0, y0) is equal to the point (x1, y1), or if the point (x1, y1) is equal to the point (x2, y2), or if both radiusX and radiusY are zero, then the method must add the point (x1, y1) to the subpath, and connect that point to the previous point (x0, y0) by a straight line.

Otherwise, if the points (x0, y0), (x1, y1), and (x2, y2) all lie on a single straight line, then the method must add the point (x1, y1) to the subpath, and connect that point to the previous point (x0, y0) by a straight line.

Otherwise, let The Arc be the shortest arc given by circumference of the ellipse that has radius radiusX on the major axis and radius radiusY on the minor axis, and whose semi-major axis is rotated rotation radians clockwise from the positive x-axis, and that has one point tangent to the half-infinite line that crosses the point (x0, y0) and ends at the point (x1, y1), and that has a different point tangent to the half-infinite line that ends at the point (x1, y1) and crosses the point (x2, y2). The points at which this ellipse touches these two lines are called the start and end tangent points respectively. The method must connect the point (x0, y0) to the start tangent point by a straight line, adding the start tangent point to the subpath, and then must connect the start tangent point to the end tangent point by The Arc, adding the end tangent point to the subpath.

The arc(x, y, radius, startAngle, endAngle, anticlockwise) and ellipse(x, y, radiusX, radiusY, rotation, startAngle, endAngle, anticlockwise) methods draw arcs.

The arc() method is equivalent to the ellipse() method in the case where the two radii are equal. When the arc() method is invoked, it must act as if the ellipse() method had been invoked with the radiusX and radiusY arguments set to the value of the radius argument, the rotation argument set to zero, and the other arguments set to the same values as their identically named arguments on the arc() method.

When the ellipse() method is invoked, it must proceed as follows. First, if the object's path has any subpaths, then the method must add a straight line from the last point in the subpath to the start point of the arc. Then, it must add the start and end points of the arc to the subpath, and connect them with an arc. The arc and its start and end points are defined as follows:

Consider an ellipse that has its origin at (x, y), that has a major-axis radius radiusX and a minor-axis radius radiusY, and that is rotated about its origin such that its semi-major axis is inclined rotation radians clockwise from the x-axis. The points at startAngle and endAngle along this circle's circumference, measured in radians clockwise from the ellipse's semi-major axis, are the start and end points respectively.

If the anticlockwise argument is false and endAngle-startAngle is equal to or greater than 2π, or, if the anticlockwise argument is true and startAngle-endAngle is equal to or greater than 2π, then the arc is the whole circumference of this ellipse.

Otherwise, the arc is the path along the circumference of this ellipse from the start point to the end point, going anti-clockwise if the anticlockwise argument is true, and clockwise otherwise. Since the points are on the ellipse, as opposed to being simply angles from zero, the arc can never cover an angle greater than 2π radians.

Negative values for radiusX or radiusY must cause the implementation to throw an IndexSizeError exception.

The rect(x, y, w, h) method must create a new subpath containing just the four points (x, y), (x+w, y), (x+w, y+h), (x, y+h), with those four points connected by straight lines, and must then mark the subpath as closed. It must then create a new subpath with the point (x, y) as the only point in the subpath.

The Path() constructor, when invoked, must return a newly created Path object.

The Path(path) constructor, when invoked, must return a newly created Path object, to which the subpaths of the argument are added. (In other words, it returns a copy of the argument.)

The Path(d) constructor must run the following steps:

1. Parse and interpret the d argument according to the SVG specification's rules for path data, thus obtaining an SVG path. [SVG]

   The resulting path could be empty. SVG defines error handling rules for parsing and applying path data.

2. Let (x, y) be the last point in the SVG path.

3. Create a new Path object and add all the subpaths in the SVG path, if any, to that Path object.

4. Create a new subpath in the Path object with (x, y) as the only point in the subpath.

5. Return the Path object as the constructed object.

The addPath(b, transform) method, when invoked on a Path object a, must run the following steps:

1. If the Path object b has no subpaths, abort these steps.

2. Create a copy of all the subpaths in b. Let this copy be known as c.

3. Transform all the coordinates and lines in c by the transform matrix transform, if it is not null.

4. Let (x, y) be the last point in the last subpath of c.

5. Add all the subpaths in c to a.

6. Create a new subpath in a with (x, y) as the only point in the subpath.

The addPathByStrokingPath(b, styles, transform) method, when invoked on a Path object a, must run the following steps:

1. If the Path object b has no subpaths, abort these steps.

2. Create a copy of all the subpaths in b. Let this copy be known as c.

3. Transform all the coordinates and lines in c by transformation matrix transform, if it is not null.

4. Let a new list of subpaths d be the result of tracing c, using the styles argument for the line styles.

5. Let (x, y) be the last point in the last subpath of d.

6. Add all the subpaths in d to a.

7. Create a new subpath in a with (x, y) as the only point in the subpath.

The addText() and addPathByStrokingText() methods each come in two variants: one rendering text at a given coordinate, and one rendering text along a given path. In both cases, the methods take a CanvasDrawingStyles object argument for the text and (if appropriate) line styles to use, an SVGMatrix object transform (which can be null), and a maximum width can optionally be provided.

When one of the addText() and addPathByStrokingText() variants that take as argument an (x, y) coordinate is invoked, the method must run the following algorithm:

1. Run the text preparation algorithm, passing it text, the CanvasDrawingStyles object argument, and, if the maxWidth argument was provided, that argument. Let glyphs be the result.

2. Move all the shapes in glyphs to the right by x CSS pixels and down by y CSS pixels.

3. Let glyph subpaths be a path describing the shapes given in glyphs, with each CSS pixel in the coordinate space of glyphs mapped to one coordinate space unit in glyph subpaths. Subpaths in glyph subpaths must wind clockwise, regardless of how the user agent's font subsystem renders fonts and regardless of how the fonts themselves are defined.

4. Transform all the coordinates and lines in glyph subpaths by the transformation matrix transform, if it is not null.

5. If the method is addPathByStrokingText(), replace glyph subpaths by the result of tracing glyph subpaths, using the CanvasDrawingStyles object argument for the line styles.

6. Let (x_final, y_final) be the last point in the last subpath of glyph subpaths.

7. Add all the subpaths in glyph subpaths to the Path object.

8. Create a new subpath in the Path object with (x_final, y_final) as the only point in the subpath.

When one of the addText() and addPathByStrokingText() variants that take as argument a Path object is invoked, the method must run the following algorithm:

1. Let target be the Path object on which the method was invoked.

2. Let path be the Path object that was provided in the method's arguments.

3. Run the text preparation algorithm, passing it text, the CanvasDrawingStyles object argument, and, if the maxWidth argument was provided, that argument. Let glyphs be the resulting array, and physical alignment be the resulting alignment value.

4. Let width be the aggregate length of all the subpaths in path, including the distances from the last point of each closed subpath to the first point of that subpath.

5. Define L to be a linear coordinate line for of all the subpaths in path, with additional lines drawn between the last point and the first point of each closed subpath, such that the first point of the first subpath is defined as point 0, and the last point of the last subpath, if the last subpath is not closed, or the second occurrence first point of that subpath, if it is closed, is defined as point width.

6. Let offset be determined according to the appropriate step below:

   If physical alignment is left

   Let offset be zero.

   If physical alignment is right

   Let offset be width.

   If physical alignment is center

   Let offset be half of width.

7. Move all the shapes in glyphs to the right by offset CSS pixels.

8. For each glyph glyph in the glyphs array, run these substeps:

   1. Let dx be the x-coordinate of the horizontal center of the bounding box of the shape described by glyph, in CSS pixels.

   2. If dx is negative or greater than width, skip the remainder of these substeps for this glyph.

   3. Recast dx to coordinate spaces units in path. (This just changes the dimensionality of dx, not its numeric value.)

   4. Find the point p on path (or implied closing lines in path) that corresponds to the position dx on the coordinate line L.

   5. Let θ be the clockwise angle from the positive x-axis to the side of the line that is tangential to path at the point p that is going in the same direction as the line at point p.

   6. Rotate the shape described by glyph clockwise by θ about the point that is at the dx coordinate horizontally and the zero coordinate vertically.

   7. Let (x, y) be the coordinate of the point p.

   8. Move the shape described by glyph to the right by x and down by y.

   9. Let glyph subpaths be a list of subpaths describing the shape given in glyph, with each CSS pixel in the coordinate space of glyph mapped to one coordinate space unit in glyph subpaths. Subpaths in glyph subpaths must wind clockwise, regardless of how the user agent's font subsystem renders fonts and regardless of how the fonts themselves are defined.

   10. Transform all the coordinates and lines in glyph subpaths by the transformation matrix transform, if it is not null.

   11. If the method is addPathByStrokingText(), replace glyph subpaths by the result of tracing glyph subpaths, using the CanvasDrawingStyles object argument for the line styles.

   12. Let (x_final, y_final) be the last point in the last subpath of glyph subpaths. (This coordinate is only used if this is the last glyph processed.)

   13. Add all the subpaths in glyph subpaths to target.

9. Create a new subpath in the Path object with (x_final, y_final) as the only point in the subpath.

The transformations must be performed in reverse order.

For instance, if a scale transformation that doubles the width is applied to the canvas, followed by a rotation transformation that rotates drawing operations by a quarter turn, and a rectangle twice as wide as it is tall is then drawn on the canvas, the actual result will be a square.

The currentTransform, on getting, must return the last object that it was set to. On setting, its value must be changed to the new value, and the transformation matrix must be updated to match the matrix described by the new value. When the CanvasRenderingContext2D object is created, the currentTransform attribute must be set a newly created SVGMatrix object. When the transformation matrix is mutated by the methods described in this section, the last SVGMatrix object to which the attribute has been set must be mutated in a corresponding fashion.

The scale(x, y) method must add the scaling transformation described by the arguments to the transformation matrix. The x argument represents the scale factor in the horizontal direction and the y argument represents the scale factor in the vertical direction. The factors are multiples.

The rotate(angle) method must add the rotation transformation described by the argument to the transformation matrix. The angle argument represents a clockwise rotation angle expressed in radians.

The translate(x, y) method must add the translation transformation described by the arguments to the transformation matrix. The x argument represents the translation distance in the horizontal direction and the y argument represents the translation distance in the vertical direction. The arguments are in coordinate space units.

The transform(a, b, c, d, e, f) method must replace the current transformation matrix with the result of multiplying the current transformation matrix with the matrix described by:

The setTransform(a, b, c, d, e, f) method must reset the current transform to the identity matrix, and then invoke the transform(a, b, c, d, e, f) method with the same arguments.

The resetTransform() method must reset the current transform to the identity matrix.

The fillStyle attribute represents the color or style to use inside shapes, and the strokeStyle attribute represents the color or style to use for the lines around the shapes.

Both attributes can be either strings, CanvasGradients, or CanvasPatterns. On setting, strings must be parsed as CSS <color> values and the color assigned, and CanvasGradient and CanvasPattern objects must be assigned themselves. [CSSCOLOR] If the value is a string but cannot be parsed as a CSS <color> value, then it must be ignored, and the attribute must retain its previous value. If the new value is a CanvasPattern object that is marked as not origin-clean, then the scratch bitmap's origin-clean flag must be set to false.

When set to a CanvasPattern or CanvasGradient object, the assignment is live, meaning that changes made to the object after the assignment do affect subsequent stroking or filling of shapes.

On getting, if the value is a color, then the serialization of the color must be returned. Otherwise, if it is not a color but a CanvasGradient or CanvasPattern, then the respective object must be returned. (Such objects are opaque and therefore only useful for assigning to other attributes or for comparison to other gradients or patterns.)

The serialization of a color for a color value is a string, computed as follows: if it has alpha equal to 1.0, then the string is a lowercase six-digit hex value, prefixed with a "#" character (U+0023 NUMBER SIGN), with the first two digits representing the red component, the next two digits representing the green component, and the last two digits representing the blue component, the digits being lowercase ASCII hex digits. Otherwise, the color value has alpha less than 1.0, and the string is the color value in the CSS rgba() functional-notation format: the literal string rgba (U+0072 U+0067 U+0062 U+0061) followed by a U+0028 LEFT PARENTHESIS, a base-ten integer in the range 0-255 representing the red component (using ASCII digits in the shortest form possible), a literal U+002C COMMA and U+0020 SPACE, an integer for the green component, a comma and a space, an integer for the blue component, another comma and space, a U+0030 DIGIT ZERO, if the alpha value is greater than zero then a U+002E FULL STOP (representing the decimal point), if the alpha value is greater than zero then one or more ASCII digits representing the fractional part of the alpha, and finally a U+0029 RIGHT PARENTHESIS. User agents must express the fractional part of the alpha value, if any, with the level of precision necessary for the alpha value, when reparsed, to be interpreted as the same alpha value.

When the context is created, the fillStyle and strokeStyle attributes must initially have the string value #000000.

When the value is a color, it must not be affected by the transformation matrix when used to draw on bitmaps.

The addColorStop(offset, color) method on the CanvasGradient interface adds a new stop to a gradient. If the offset is less than 0 or greater than 1 then an IndexSizeError exception must be thrown. If the color cannot be parsed as a CSS <color> value, then a SyntaxError exception must be thrown. Otherwise, the gradient must have a new stop placed, at offset offset relative to the whole gradient, and with the color obtained by parsing color as a CSS <color> value. If multiple stops are added at the same offset on a gradient, they must be placed in the order added, with the first one closest to the start of the gradient, and each subsequent one infinitesimally further along towards the end point (in effect causing all but the first and last stop added at each point to be ignored).

The createLinearGradient(x0, y0, x1, y1) method takes four arguments that represent the start point (x0, y0) and end point (x1, y1) of the gradient. The method must return a linear CanvasGradient initialized with the specified line.

Linear gradients must be rendered such that all points on a line perpendicular to the line that crosses the start and end points have the color at the point where those two lines cross (with the colors coming from the interpolation and extrapolation described above). The points in the linear gradient must be transformed as described by the current transformation matrix when rendering.

If x0 = x1 and y0 = y1, then the linear gradient must paint nothing.

The createRadialGradient(x0, y0, r0, x1, y1, r1) method takes six arguments, the first three representing the start circle with origin (x0, y0) and radius r0, and the last three representing the end circle with origin (x1, y1) and radius r1. The values are in coordinate space units. If either of r0 or r1 are negative, an IndexSizeError exception must be thrown. Otherwise, the method must return a radial CanvasGradient initialized with the two specified circles.

Radial gradients must be rendered by following these steps:

1. If x₀ = x₁ and y₀ = y₁ and r₀ = r₁, then the radial gradient must paint nothing. Abort these steps.

2. Let x(ω) = (x₁-x₀)ω + x₀

   Let y(ω) = (y₁-y₀)ω + y₀

   Let r(ω) = (r₁-r₀)ω + r₀

   Let the color at ω be the color at that position on the gradient (with the colors coming from the interpolation and extrapolation described above).

3. For all values of ω where r(ω) > 0, starting with the value of ω nearest to positive infinity and ending with the value of ω nearest to negative infinity, draw the circumference of the circle with radius r(ω) at position (x(ω), y(ω)), with the color at ω, but only painting on the parts of the bitmap that have not yet been painted on by earlier circles in this step for this rendering of the gradient.

This effectively creates a cone, touched by the two circles defined in the creation of the gradient, with the part of the cone before the start circle (0.0) using the color of the first offset, the part of the cone after the end circle (1.0) using the color of the last offset, and areas outside the cone untouched by the gradient (transparent black).

The resulting radial gradient must then be transformed as described by the current transformation matrix when rendering.

Gradients must be painted only where the relevant stroking or filling effects requires that they be drawn.

To create objects of this type, the createPattern(image, repetition) method is used. When the method is invoked, the user agent must run the following steps:

1. Let image be the first argument and repetition be the second argument.

2. Check the usability of the image argument. If this returns aborted, then an exception has been thrown and the method doesn't return anything; abort these steps. If it returns bad, then return null and abort these steps. Otherwise it returns good; continue with these steps.

3. If repetition is the empty string, let it be "repeat".

4. If repetition is not a case-sensitive match for one of "repeat", "repeat-x", "repeat-y", or "no-repeat", throw a SyntaxError exception and abort these steps.

5. Create a new CanvasPattern object with the image image and the repetition behavior given by repetition.

6. If the image argument is not origin-clean, then mark the CanvasPattern object as not origin-clean.

7. Return the CanvasPattern object.

Modifying the image used when creating a CanvasPattern object after calling the createPattern() method must not affect the pattern(s) rendered by the CanvasPattern object.

Patterns have a transformation matrix, which controls how the pattern is used when it is painted. Initially, a pattern's transformation matrix must be the identity transform.

When the setTransform() method is invoked on the pattern, the user agent must replace the pattern's transformation matrix with the one described by the SVGMatrix object provided as an argument to the method.

When a pattern is to be rendered within an area, the user agent must run the following steps to determine what is rendered:

1. Create an infinite transparent black bitmap.

2. Place a copy of the image on the bitmap, anchored such that its top left corner is at the origin of the coordinate space, with one coordinate space unit per CSS pixel of the image, then place repeated copies of this image horizontally to the left and right, if the repetition behavior is "repeat-x", or vertically up and down, if the repetition behavior is "repeat-y", or in all four directions all over the bitmap, if the repetition behavior is "repeat".

   If the original image data is a bitmap image, the value painted at a point in the area of the repetitions is computed by filtering the original image data. When scaling up, if the imageSmoothingEnabled attribute is set to true, the image must be rendered using nearest-neighbor interpolation. Otherwise, the user agent may use any filtering algorithm (for example bilinear interpolation or nearest-neighbor). When such a filtering algorithm requires a pixel value from outside the original image data, it must instead use the value from wrapping the pixel's coordinates to the original image's dimensions. (That is, the filter uses 'repeat' behavior, regardless of the value of the pattern's repetition behavior.)

3. Transform the resulting bitmap according to the pattern's transformation matrix.

4. Transform the resulting bitmap again, this time according to the current transformation matrix.

5. Replace any part of the image outside the area in which the pattern is to be rendered with transparent black.

6. The resulting bitmap is what is to be rendered, with the same origin and same scale.

If a radial gradient or repeated pattern is used when the transformation matrix is singular, the resulting style must be transparent black (otherwise the gradient or pattern would be collapsed to a point or line, leaving the other pixels undefined). Linear gradients and solid colors always define all points even with singular tranformation matrices.

The current transformation matrix must be applied to the following four coordinates, which form the path that must then be closed to get the specified rectangle: (x, y), (x+w, y), (x+w, y+h), (x, y+h).

Shapes are painted without affecting the current default path, and are subject to the clipping region, and, with the exception of clearRect(), also shadow effects, global alpha, and global composition operators.

The clearRect(x, y, w, h) method must run the following steps:

1. Let pixels be the set of pixels in the specified rectangle that also intersect the current clipping region.

2. Clear the pixels in pixels to a fully transparent black, erasing any previous image.

3. Clear regions that cover the pixels in pixels on the scratch bitmap.

If either height or width are zero, this method has no effect, since the set of pixels would be empty.

The fillRect(x, y, w, h) method must paint the specified rectangular area using the fillStyle. If either height or width are zero, this method has no effect.

The strokeRect(x, y, w, h) method must take the result of tracing the path described below, using the CanvasRenderingContext2D object's line styles, and fill it with the strokeStyle.

If both w and h are zero, the path has a single subpath with just one point (x, y), and no lines, and this method thus has no effect (the trace a path algorithm returns an empty path in that case).

If just one of either w or h is zero, then the path has a single subpath consisting of two points, with coordinates (x, y) and (x+w, y+h), in that order, connected by a single straight line.

Otherwise, the path has a single subpath consisting of four points, with coordinates (x, y), (x+w, y), (x+w, y+h), and (x, y+h), connected to each other in that order by straight lines.

The CanvasRenderingContext2D interface provides the following methods for rendering text.

The fillText() and strokeText() methods take three or four arguments, text, x, y, and optionally maxWidth, and render the given text at the given (x, y) coordinates ensuring that the text isn't wider than maxWidth if specified, using the current font, textAlign, and textBaseline values. Specifically, when the methods are called, the user agent must run the following steps:

1. Run the text preparation algorithm, passing it text, the CanvasRenderingContext2D object, and, if the maxWidth argument was provided, that argument. Let glyphs be the result.

2. Move all the shapes in glyphs to the right by x CSS pixels and down by y CSS pixels.

3. Paint the shapes given in glyphs, as transformed by the current transformation matrix, with each CSS pixel in the coordinate space of glyphs mapped to one coordinate space unit.

   For fillText(), fillStyle must be applied to the shapes and strokeStyle must be ignored. For strokeText(), the reverse holds: strokeStyle must be applied to the result of tracing the shapes using the CanvasRenderingContext2D object for the line styles, and fillStyle must be ignored.

   These shapes are painted without affecting the current path, and are subject to shadow effects, global alpha, the clipping region, and global composition operators.

4. If the text preparation algorithm used a font that has an origin that is not the same as the entry script's origin (even if "using a font" means just checking if that font has a particular glyph in it before falling back to another font), then set the scratch bitmap's origin-clean flag to false.

The measureText() method takes one argument, text. When the method is invoked, the user agent must run the text preparation algorithm, pass a new TextMetrics object with its attributes set as described in the following list. If doing these measurements requires using a font that has an origin that is not the same as that of the Document object that owns the canvas element (even if "using a font" means just checking if that font has a particular glyph in it before falling back to another font), then the method must throw a SecurityError exception. Otherwise, it must return the new TextMetrics object. [CSS]

The beginPath() method must empty the list of subpaths in the context's current default path so that the it once again has zero subpaths.

Where the following method definitions use the term intended path, it means the Path argument, if one was provided, or the current default path otherwise.

When the intended path is a Path object, the coordinates and lines of its subpaths must be transformed according to the CanvasRenderingContext2D object's current transformation matrix when used by these methods (without affecting the Path object itself). When the intended path is the current default path, it is not affected by the transform. (This is because transformations already affect the current default path when it is constructed, so applying it when it is painted as well would result in a double transformation.)

The fill() method must fill all the subpaths of the intended path, using fillStyle, and using the fill rule indicated by the fillRule argument. Open subpaths must be implicitly closed when being filled (without affecting the actual subpaths).

The stroke() method must trace the intended path, using the CanvasRenderingContext2D object for the line styles, and then fill the resulting path using the strokeStyle attribute, using the non-zero winding rule.

As a result of how the algorithm to trace a path is defined, overlapping parts of the paths in one stroke operation are treated as if their union was what was painted.

The stroke style is affected by the transformation during painting, even if the intended path is the current default path.

Paths, when filled or stroked, must be painted without affecting the current default path or any Path objects, and must be subject to shadow effects, global alpha, the clipping region, and global composition operators. (The effect of transformations is described above and varies based on which path is being used.)

Zero-length line segments must be pruned before stroking a path. Subpaths with just one point must be ignored.

The drawSystemFocusRing(element) method, when invoked, must run the following steps:

1. If element is not focused or is not a descendant of the element with whose context the method is associated, then abort these steps.

2. If the user has requested the use of particular focus rings (e.g. high-contrast focus rings), or if the element would have a focus ring drawn around it, then draw a focus ring of the appropriate style along the intended path, following platform conventions.

   Some platforms only draw focus rings around elements that have been focused from the keyboard, and not those focused from the mouse. Other platforms simply don't draw focus rings around some elements at all unless relevant accessibility features are enabled. This API is intended to follow these conventions. User agents that implement distinctions based on the manner in which the element was focused are encouraged to classify focus driven by the focus() method based on the kind of user interaction event from which the call was triggered (if any).

   The focus ring should not be subject to the shadow effects, the global alpha, the global composition operators, the fillStyle attribute, the strokeStyle attribute, or any of the CanvasDrawingStyles members, but should be subject to the clipping region. (The effect of transformations is described above and varies based on which path is being used.)

3. Optionally, run the appropriate step from the following list:

   If the CanvasRenderingContext2D object's context bitmap mode is fixed

   Inform the user that the focus is at the location given by the intended path. User agents may wait until the next time the event loop reaches its "update the rendering" step to optionally inform the user.

   Otherwise

   Add instructions to the scratch bitmap's list of pending interface actions that inform the user that the focus is at the location of the bitmap given by the intended path.

The drawCustomFocusRing(element) method, when invoked, must run the following steps:

1. If element is not focused or is not a descendant of the element with whose context the method is associated, then return false and abort these steps.

2. Let result be true.

3. If the user has requested the use of particular focus rings (e.g. high-contrast focus rings), then draw a focus ring of the appropriate style along the intended path, and set result to false.

   The focus ring should not be subject to the shadow effects, the global alpha, the global composition operators, the fillStyle attribute, the strokeStyle attribute, or any of the CanvasDrawingStyles members, but should be subject to the clipping region. (The effect of transformations is described above and varies based on which path is being used.)

4. Optionally, run the appropriate step from the following list:

   If the CanvasRenderingContext2D object's context bitmap mode is fixed

   Inform the user that the focus is at the location given by the intended path. The user agent may wait until the next time the event loop reaches its "update the rendering" step to optionally inform the user.

   Otherwise

   Add instructions to the scratch bitmap's list of pending interface actions that inform the user that the focus is at the location of the bitmap given by the intended path.

5. Return result.

User agents should not implicitly close open subpaths in the intended path when drawing the focus ring.

This might be a moot point, however. For example, if the focus ring is drawn as an axis-aligned bounding rectangle around the points in the intended path, then whether the subpaths are closed or not has no effect. This specification intentionally does not specify precisely how focus rings are to be drawn: user agents are expected to honor their platform's native conventions.

The scrollPathIntoView() method, when invoked, if the CanvasRenderingContext2D object's context bitmap mode is fixed, must run the following steps; and otherwise, must add instructions to the scratch bitmap's list of pending interface actions that run the following steps:

1. Let the specified rectangle be the rectangle of the bounding box of the intended path.

2. Let notional child be a hypothetical element that is a rendered child of the canvas element whose dimensions are those of the specified rectangle.

3. Scroll notional child into view with the align to top flag set.

4. Optionally, inform the user that the caret or selection (or both) cover the specified rectangle of the canvas. If the CanvasRenderingContext2D object's context bitmap mode was fixed when the method was invoked, the user agent may wait until the next time the event loop reaches its "update the rendering" step to optionally inform the user.

"Inform the user", as used in this section, could mean calling a system accessibility API, which would notify assistive technologies such as magnification tools. To properly drive magnification based on a focus change, a system accessibility API driving a screen magnifier needs the bounds for the newly focused object. The methods above are intended to enable this by allowing the user agent to report the bounding box of the path used to render the focus ring as the bounds of the element element passed as an argument, if that element is focused, and the bounding box of the area to which the user agent is scrolling as the bounding box of the current selection.

The clip() method must create a new clipping region by calculating the intersection of the current clipping region and the area described by the intended path, using the fill rule indicated by the fillRule argument. Open subpaths must be implicitly closed when computing the clipping region, without affecting the actual subpaths. The new clipping region replaces the current clipping region.

When the context is initialized, the clipping region must be set to the largest infinite surface (i.e. by default, no clipping occurs).

The resetClip() method must create a new clipping region that is the largest infinite surface. The new clipping region replaces the current clipping region.

The isPointInPath() method must return true if the point given by the x and y coordinates passed to the method, when treated as coordinates in the canvas coordinate space unaffected by the current transformation, is inside the intended path as determined by the fill rule indicated by the fillRule argument; and must return false otherwise. Points on the path itself must be considered to be inside the path. If either of the arguments is infinite or NaN, then the method must return false.

The isPointInStroke() method must return true if the point given by the x and y coordinates passed to the method, when treated as coordinates in the canvas coordinate space unaffected by the current transformation, is inside the path that results from tracing the intended path, using the non-zero winding rule, and using the CanvasRenderingContext2D object for the line styles; and must return false otherwise. Points on the resulting path must be considered to be inside the path. If either of the arguments is infinite or NaN, then the method must return false.

This method can be invoked with three different sets of arguments:

• drawImage(image, dx, dy)
• drawImage(image, dx, dy, dw, dh)
• drawImage(image, sx, sy, sw, sh, dx, dy, dw, dh)

When the drawImage() method is invoked, the user agent must run the following steps:

1. Check the usability of the image argument. If this returns aborted, then an exception has been thrown and the method doesn't return anything; abort these steps. If it returns bad, then abort these steps without drawing anything. Otherwise it returns good; continue with these steps.

2. Establish the source and destination rectangles as follows:

   If not specified, the dw and dh arguments must default to the values of sw and sh, interpreted such that one CSS pixel in the image is treated as one unit in the scratch bitmap's coordinate space. If the sx, sy, sw, and sh arguments are omitted, they must default to 0, 0, the image's intrinsic width in image pixels, and the image's intrinsic height in image pixels, respectively. If the image has no intrinsic dimensions, the concrete object size must be used instead, as determined using the CSS "Concrete Object Size Resolution" algorithm, with the specified size having neither a definite width nor height, nor any additional contraints, the object's intrinsic properties being those of the image argument, and the default object size being the size of the scratch bitmap. [CSSIMAGES]

   The source rectangle is the rectangle whose corners are the four points (sx, sy), (sx+sw, sy), (sx+sw, sy+sh), (sx, sy+sh).

   The destination rectangle is the rectangle whose corners are the four points (dx, dy), (dx+dw, dy), (dx+dw, dy+dh), (dx, dy+dh).

   When the source rectangle is outside the source image, the source rectangle must be clipped to the source image and the destination rectangle must be clipped in the same proportion.

   When the destination rectangle is outside the destination image (the scratch bitmap), the pixels that land outside the scratch bitmap are discarded, as if the destination was an infinite canvas whose rendering was clipped to the dimensions of the scratch bitmap.

3. If one of the sw or sh arguments is zero, abort these steps. Nothing is painted.

4. Paint the region of the image argument specified by the source rectangle on the region of the rendering context's scratch bitmap specified by the destination rectangle, after applying the current transformation matrix to the destination rectangle.

   The image data must be processed in the original direction, even if the dimensions given are negative.

   When scaling up, if the imageSmoothingEnabled attribute is set to true, the user agent should attempt to apply a smoothing algorithm to the image data when it is scaled. Otherwise, the image must be rendered using nearest-neighbor interpolation.

   This specification does not define the precise algorithm to use when scaling an image down, or when scaling an image up when the imageSmoothingEnabled attribute is set to true.

   When a canvas or CanvasRenderingContext2D object is drawn onto itself, the drawing model requires the source to be copied before the image is drawn, so it is possible to copy parts of a canvas or scratch bitmap onto overlapping parts of itself.

   If the original image data is a bitmap image, the value painted at a point in the destination rectangle is computed by filtering the original image data. The user agent may use any filtering algorithm (for example bilinear interpolation or nearest-neighbor). When the filtering algorithm requires a pixel value from outside the original image data, it must instead use the value from the nearest edge pixel. (That is, the filter uses 'clamp-to-edge' behavior.) When the filtering algorithm requires a pixel value from outside the source rectangle but inside the original image data, then the value from the original image data must be used.

   Thus, scaling an image in parts or in whole will have the same effect. This does mean that when sprites coming from a single sprite sheet are to be scaled, adjacent images in the sprite sheet can interfere. This can be avoided by ensuring each sprite in the sheet is surrounded by a border of transparent black, or by copying sprites to be scaled into temporary canvas elements and drawing the scaled sprites from there.

   Images are painted without affecting the current path, and are subject to shadow effects, global alpha, the clipping region, and global composition operators.

5. If the image argument is not origin-clean, set the scratch bitmap's origin-clean flag to false.

A hit region A is an ancestor region of a hit region B if B has a parent and its parent is either A or another hit region for which A is an ancestor region.

The region identified by the ID ID in a bitmap bitmap is the value returned by the following algorithm (which can return a hit region or nothing):

1. If ID is null, return nothing and abort these steps.

2. Let list be the hit region list associated with bitmap.

3. If there is a hit region in list whose ID is a case-sensitive match for ID, then return that hit region and abort these steps.

4. Otherwise, return nothing.

The region representing the control control for a bitmap bitmap is the value returned by the following algorithm (which can return a hit region or nothing):

1. Let list be the hit region list associated with bitmap.

2. If there is a hit region in list whose control is control, then return that hit region and abort these steps.

3. Otherwise, return nothing.

The control represented by a region region for a canvas element ancestor is the value returned by the following algorithm (which can return an element or nothing):

1. If region has no control, return nothing and abort these steps.

2. Let control be region's control.

3. If control is not a descendant of ancestor, then return nothing and abort these steps.

4. Otherwise, return control.

The cursor for a hit region region of a canvas element ancestor is the value returned by the following algorithm:

1. Loop: If region has a cursor specification other than "inherit", then return that hit region's cursor specification and abort these steps.

2. If region has a parent, then let region be that hit region's parent, and return to the step labeled loop.

3. Otherwise, return the used value of the 'cursor' property for the canvas element, if any; if there isn't one, return 'auto'. [CSSUI]

The region for a pixel pixel on a bitmap bitmap is the value returned by the following algorithm (which can return a hit region or nothing):

1. Let list be the hit region list associated with bitmap.

2. If there is a hit region in list whose set of pixels contains pixel, then return that hit region and abort these steps.

3. Otherwise, return nothing.

To clear regions that cover the pixels pixels on a bitmap bitmap, the user agent must run the following steps:

1. Let list be the hit region list associated with bitmap.

2. Remove all pixels in pixels from the set of pixels of each hit region in list.

3. Garbage-collect the regions of bitmap.

To garbage-collect the regions of a bitmap bitmap, the user agent must run the following steps:

1. Let list be the hit region list associated with bitmap.

2. Loop: Let victim be the first hit region in list to have an empty set of pixels and a zero child count, if any. If there is no such hit region, abort these steps.

3. If victim has a parent, then decrement that hit region's child count by one.

4. Remove victim from list.

5. Jump back to the step labeled loop.

Adding a new region and calling clearRect() are the two ways this clearing algorithm can be invoked. The hit region list itself is also reset when the rendering context is reset, e.g. when a CanvasRenderingContext2D object is bound to or unbound from a canvas, or when the dimensions of the bitmap are changed.

When the addHitRegion() method is invoked, the user agent must run the following steps:

1. Let arguments be the dictionary object provided as the method's argument.

2. If the arguments object's path member is not null, let source path be the path member's value. Otherwise, let it be the CanvasRenderingContext2D object's current default path.

3. Transform all the coordinates and lines in source path by the current transform matrix, if the arguments object's path member is not null.

4. Let specified pixels be the pixels contained in source path, using the fill rule indicated by the fillRule member.

5. If the arguments object's id member is the empty string, let it be null instead.

6. If the arguments object's id member is not null, then let previous region for this ID be the region identified by the ID given by the id member's value in this scratch bitmap, if any. If the id member is null or no such region currently exists, let previous region for this ID be null.

7. If the arguments object's parent member is the empty string, let it be null instead.

8. If the arguments object's parent member is not null, then let parent region be the region identified by the ID given by the parent member's value in the scratch bitmap, if any. If the parent member is null or no such region currently exists, let parent region be null.

9. If the arguments object's label member is the empty string, let it be null instead.

10. If any of the following conditions are met, throw a NotSupportedError exception and abort these steps.

    • The arguments object's control and label members are both non-null.
    • The arguments object's control and role members are both non-null.
    • The arguments object's role member's value is the empty string, and the label member's value is either null or the empty string.
    • The specified pixels has no pixels.
    • The arguments object's control member is not null but is neither an a element that represents a hyperlink, a button element, an input element whose type attribute is in one of the Checkbox or Radio Button states, nor an input element that is a button.
    • The parent region is not null but has a control.
    • The previous region for this ID is the same hit region as the parent region.
    • The previous region for this ID is an ancestor region of the parent region.

11. If the parent member is not null but parent region is null, then throw a NotFoundError exception and abort these steps.

12. If any of the following conditions are met, throw a SyntaxError exception and abort these steps.

    • The arguments object's cursor member is not null but is neither an ASCII case-insensitive match for the string "inherit", nor a valid CSS 'cursor' property value. [CSSUI]
    • The arguments object's role member is not null but its value is not an ordered set of unique space-separated tokens whose tokens are all case-sensitive matches for names of non-abstract WAI-ARIA roles. [ARIA]

13. Let region be a newly created hit region, with its information configured as follows:

    Hit region's set of pixels

    The specified pixels

    Hit region's bounding circumference

    A user-agent-defined shape that wraps the pixels contained in source path. (In the simplest case, this can just be the bounding rectangle; this specification allows it to be any shape in order to allow other interfaces.)

    Hit region's ID

    If the arguments object's id member is not null: the value of the id member. Otherwise, region has no id.

    Hit region's parent

    If parent region is not null: parent region. Otherwise, region has no parent.

    Hit region's child count

    Initially zero.

    Hit region's cursor specification

    If parent region is not null: parent region. Otherwise, region has no parent.

    Hit region's control

    If the arguments object's control member is not null: the value of the control member. Otherwise, region has no control.

    Hit region's label

    If the arguments object's label member is not null: the value of the label member. Otherwise, region has no label.

    Hit region's ARIA role

    If the arguments object's role member is not null: the value of the role member (which might be the empty string). Otherwise, if the arguments object's label member is not null: the empty string. Otherwise, region has no ARIA role.

14. If the arguments object's cursor member is not null, then act as if a CSS rule for the canvas element setting its 'cursor' property had been seen, whose value was the hit region's cursor specification.

    For example, if the user agent prefetches cursor values, this would cause that to happen in response to an appropriately-formed addHitRegion() call.

15. If the arguments object's control member is not null, then let previous region for the control be the region representing the control given by the control member's value for this scratch bitmap, if any. If the control member is null or no such region currently exists, let previous region for the control be null.

16. If there is a previous region with this control, remove it from the scratch bitmap's hit region list; then, if it had a parent region, decrement that hit region's child count by one.

17. If there is a previous region with this ID, remove it, and all hit regions for which it is an ancestor region, from the scratch bitmap's hit region list; then, if it had a parent region, decrement that hit region's child count by one.

18. If there is a parent region, increment its hit region's child count by one.

19. Clear regions that cover the pixels in region's set of pixels on this scratch bitmap.

20. Add region to the scratch bitmap's element's hit region list.

If the hitregion is interactive or requires the use of WAI-ARIA states or properties (in addition to role or label), authors must not use an unbacked region description (as there is no method to attach WAI-ARIA states or properties to an unbacked region description and unbacked region are unable to receive focus and therefore cannot be in a focused state).

In cases where the region is functioning as an interactive control it must have a control. It is recommended that the control() references an html interactive element unless html does not provide the desired features. In which case WAI-ARIA role, states and properties may be used as allowed in html.

Examples of regions where a WAI-ARIA role requires keyboard focus or the use of WAI-ARIA states and properties other than label, are a button and a checkbox. A button has a role="button" and it is required to be focusable using the keyboard. A checkbox has a role ="checkbox", requires an aria-checked state and it is required to be focusable using the keyboard. [ARIA]

When the removeHitRegion() method is invoked, the user agent must run the following steps:

1. Let region be the region identified by the ID given by the method's argument in the rendering context's scratch bitmap. If no such region currently exists, abort these steps.

   If the method's argument is the empty string, then no region will match.

2. Remove region, and all hit regions for which it is an ancestor region, from the rendering context's scratch bitmap's hit region list; then, if it had a parent region, decrement that hit region's child count by one.

3. Garbage-collect the regions of the rendering context's scratch bitmap.

The region attribute on MouseEvent objects must return the value it was initialized to. When the object is created, this attribute must be initialized to null. It represents the hit region's ID if the mouse was over a hit region when the event was fired.

When a MouseEvent is to be fired at a canvas element by the user agent in response to a pointing device action, if the canvas element has a hit region list, the user agent must instead follow these steps. If these steps say to act as normal, that means that the event must be fired as it would have had these requirements not been applied.

1. If the pointing device is not indicating a pixel on the canvas, act as normal and abort these steps.

2. Let pixel be the pixel indicated by the pointing device.

3. Let region be the hit region that is the region for the pixel pixel on this canvas element's bitmap, if any.

4. If there is no region, then act as normal and abort these steps.

5. Let id be the region's ID, if any.

6. If there is an id, then initialize the event object's region attribute to id.

7. Let control be the control represented by region for this canvas element, if any.

8. If there is a control, then target the event object at control instead of the canvas element.

9. Continue dispatching the event, but with the updated event object and target as given in the above steps.

When a user's pointing device cursor is positioned over a canvas element, user agents should render the pointing device cursor according to the cursor specification described by the cursor for the hit region that is the region for the pixel that the pointing device designates on the canvas element's bitmap.

User agents are encouraged to make use of the information present in a canvas element's hit region list to improve the accessibility of canvas elements.

Each hit region should be handled in a fashion equivalent to a node in a virtual DOM tree rooted at the canvas element. The hierarchy of this virtual DOM tree must match the hierarchy of the hit regions, as described by the parent of each region. Regions without a parent must be treated as children of the canvas element for the purpose of this virtual DOM tree. For each node in such a DOM tree, the hit region's bounding circumference gives the region of the screen to use when representing the node (if appropriate).

The semantics of a hit region for the purposes of this virtual DOM tree are those of the the control represented by the region, if it has one, or else of a non-interactive element whose ARIA role, if any, is that given by the hit region's ARIA role, and whose textual representation, if any, is given by the hit region's label.

For the purposes of accessibility tools, when an element C is a descendant of a canvas element and there is a region representing the control C for that canvas element's bitmap, then the element's position relative to the document should be presented as if it was that region in the canvas element's virtual DOM tree.

The semantics of a hit region for the purposes of this virtual DOM tree are those of the the control represented by the region, if it has one, or else of a non-interactive element whose ARIA role, if any, is that given by the hit region's ARIA role, and whose textual representation, if any, is given by the hit region's label.

Thus, for instance, a user agent on a touch-screen device could provide haptic feedback when the user croses over a hit region's bounding circumference, and then read the hit region's label to the user. Similarly, a desktop user agent with a virtual accessibility focus separate from the keyboard input focus could allow the user to navigate through the hit regions, using the virtual DOM tree described above to enable hierarchical navigation. When an interactive control inside the canvas element is focused, if the control has a corresponding region, then that hit region's bounding circumference could be used to determine what area of the display to magnify.

The createImageData() and createImageDataHD() methods are used to instantiate new blank ImageData objects.

When the createImageData() method is invoked with two arguments sw and sh, it must return a new ImageData object representing a rectangle with a width in equal to the absolute magnitude of sw and a height equal to the absolute magnitude of sh, and with a 1.0 pixel density, if both sw and sh are non-zero. If one or both of sw and sh are zero, then the method must throw an IndexSizeError exception instead. When invoked with a single imagedata argument, it must return a new ImageData object representing a rectangle with the same dimensions and pixel density as the ImageData object passed as the argument. When the method returns an ImageData object, it must be filled with transparent black.

When the createImageDataHD() method is invoked and both of its arguments (sw and sh) are non-zero, it must return a new ImageData object representing a rectangle with a width equal to the absolute magnitude of sw multiplied by scale, a height equal to the absolute magnitude of sh multiplied by scale, and a pixel density equal to scale, where scale is the number of pixels in the scratch bitmap per coordinate space units. The ImageData object returned must be filled with transparent black. If either or both of the arguments are zero, the method must instead throw an IndexSizeError exception.

The getImageData(sx, sy, sw, sh) method must, if either the sw or sh arguments are zero, throw an IndexSizeError exception; otherwise, if the scratch bitmap's origin-clean flag is set to false, it must throw a SecurityError exception; otherwise, it must return an ImageData object with width sw and height sh representing the scratch bitmap for the area of that bitmap denoted by the rectangle whose corners are the four points (sx, sy), (sx+sw, sy), (sx+sw, sy+sh), (sx, sy+sh), in the bitmap's coordinate space units. If the bitmap does not represent each coordinate space unit square using exactly one pixel, the value of each pixel in the returned object must be derived from the value(s) of the pixel(s) in the bitmap that correspond to the same coordinate. Pixels outside the scratch bitmap must be returned as transparent black. Pixels must be returned as non-premultiplied alpha values. The pixel density of the object returned must be 1.0.

The getImageDataHD(sx, sy, sw, sh) method must, if either the sw or sh arguments are zero, throw an IndexSizeError exception; otherwise, if the scratch bitmap's origin-clean flag is set to false, it must throw a SecurityError exception; otherwise, it must return an ImageData object with width sw multiplied by scale and height sh multiplied by scale representing the scratch bitmap for the area of that bitmap denoted by the rectangle whose corners are the four points (sx, sy), (sx+sw, sy), (sx+sw, sy+sh), (sx, sy+sh), in the bitmap's coordinate space units. Pixels outside the scratch bitmap must be returned as transparent black. Pixels must be returned as non-premultiplied alpha values. The pixel density of the object returned must be scale. For the purposes of this paragraph, scale is the number of pixels in the scratch bitmap per coordinate space units.

New ImageData objects must be initialized so that their width attribute is set to the number of pixels per row in the image data, their height attribute is set to the number of rows in the image data, their resolution is set to the object's pixel density, and their data attribute is initialized to a Uint8ClampedArray object. The Uint8ClampedArray object must use a Canvas Pixel ArrayBuffer for its storage, and must have a zero start offset and a length equal to the length of its storage, in bytes. The Canvas Pixel ArrayBuffer must contain the image data. At least one pixel's worth of image data must be returned. [TYPEDARRAY]

A Canvas Pixel ArrayBuffer is an ArrayBuffer that whose data is represented in left-to-right order, row by row top to bottom, starting with the top left, with each pixel's red, green, blue, and alpha components being given in that order for each pixel. Each component of each pixel represented in this array must be in the range 0..255, representing the 8 bit value for that component. The components must be assigned consecutive indices starting with 0 for the top left pixel's red component. [TYPEDARRAY]

The putImageData() and putImageDataHD() methods write data from ImageData structures back to the rendering context's scratch bitmap. Their arguments are: imagedata, dx, dy, dirtyX, dirtyY, dirtyWidth, and dirtyHeight.

When the last four arguments to these methods are omitted, they must be assumed to have the values 0, 0, the width member of the imagedata structure, and the height member of the imagedata structure, respectively.

When invoked, these methods must act as follows:

1. If dirtyWidth is negative, let dirtyX be dirtyX+dirtyWidth, and let dirtyWidth be equal to the absolute magnitude of dirtyWidth.

   If dirtyHeight is negative, let dirtyY be dirtyY+dirtyHeight, and let dirtyHeight be equal to the absolute magnitude of dirtyHeight.

2. If dirtyX is negative, let dirtyWidth be dirtyWidth+dirtyX, and let dirtyX be zero.

   If dirtyY is negative, let dirtyHeight be dirtyHeight+dirtyY, and let dirtyY be zero.

3. If dirtyX+dirtyWidth is greater than the width attribute of the imagedata argument, let dirtyWidth be the value of that width attribute, minus the value of dirtyX.

   If dirtyY+dirtyHeight is greater than the height attribute of the imagedata argument, let dirtyHeight be the value of that height attribute, minus the value of dirtyY.

4. If, after those changes, either dirtyWidth or dirtyHeight is negative or zero, stop these steps without affecting any bitmaps.

5. Run the appropriate steps from the following list:

   If the method was putImageData()

   Draw the region of the image data in the horizontal rectangle whose top left corner is at (dirtyX,dirtyY) and whose bottom right corner is at (dirtyX+dirtyWidth,dirtyY+dirtyHeight) onto the rendering context's scratch bitmap, aligned such that the top left of the rectangle is at coordinate (dx,dy).

   If the imageSmoothingEnabled attribute is set to true, then the user agent should attempt to apply a smoothing algorithm to the image data if the scratch bitmap does not have exactly one device pixel per coordinate space unit.

   If the method was putImageDataHD()

   1. Let dx_device be the x-coordinate of the device pixel in the scratch bitmap corresponding to the dx coordinate in the scratch bitmap's coordinate space.

      Let dy_device be the y-coordinate of the device pixel in the scratch bitmap corresponding to the dy coordinate in the scratch bitmap's coordinate space.

   2. For all integer values of x and y where dirtyX ≤ x < dirtyX+dirtyWidth and dirtyY ≤ y < dirtyY+dirtyHeight, copy the four channels of the pixel with coordinate (x, y) in the imagedata data structure to the pixel with coordinate (dx_device+x, dy_device+y) in the rendering context's scratch bitmap.

The handling of pixel rounding when the specified coordinates do not exactly map to the device coordinate space is not defined by this specification, except that the following must result in no visible changes to the rendering:

context.putImageDataHD(context.getImageDataHD(x, y, w, h), p, q);

...for any value of x, y, w, and h and where p is the smaller of x and the sum of x and w, and q is the smaller of y and the sum of y and h; and except that the following two calls:

context.createImageData(w, h);
context.getImageData(0, 0, w, h);

...must return ImageData objects with the same dimensions as each other, and the following two calls:

context.createImageDataHD(w, h);
context.getImageDataHD(0, 0, w, h);

...must also return ImageData objects with the same dimensions as each other, for any value of w and h in both cases. In other words, while user agents may round the arguments of these methods so that they map to device pixel boundaries, any rounding performed must be performed consistently for all of the methods described in this section.

Due to the lossy nature of converting to and from premultiplied alpha color values, pixels that have just been set using putImageDataHD() might be returned to an equivalent getImageDataHD() as different values.

The current path, transformation matrix, shadow attributes, global alpha, the clipping region, and global composition operator must not affect the methods described in this section.

All drawing operations are affected by the global compositing attributes, globalAlpha and globalCompositeOperation.

The globalAlpha attribute gives an alpha value that is applied to shapes and images before they are composited onto the scratch bitmap. The value must be in the range from 0.0 (fully transparent) to 1.0 (no additional transparency). If an attempt is made to set the attribute to a value outside this range, including Infinity and Not-a-Number (NaN) values, the attribute must retain its previous value. When the context is created, the globalAlpha attribute must initially have the value 1.0.

The globalCompositeOperation attribute sets the current composition operator, which controls how shapes and images are drawn onto the scratch bitmap, once they have had globalAlpha and the current transformation matrix applied. The possible values are those defined in the Compositing and Blending specification. [COMPOSITE]

These values are all case-sensitive — they must be used exactly as defined. User agents must not recognize values that are not a case-sensitive match for one of the values given in the Compositing and Blending specification. [COMPOSITE]

On setting, if the user agent does not recognize the specified value, it must be ignored, leaving the value of globalCompositeOperation unaffected. Otherwise, the attribute must be set to the given new value.

When the context is created, the globalCompositeOperation attribute must initially have the value source-over.

The imageSmoothingEnabled attribute, on getting, must return the last value it was set to. On setting, it must be set to the new value. When the CanvasRenderingContext2D object is created, the attribute must be set to true.

The shadowColor attribute sets the color of the shadow.

When the context is created, the shadowColor attribute initially must be fully-transparent black.

On getting, the serialization of the color must be returned.

On setting, the new value must be parsed as a CSS <color> value and the color assigned. If the value cannot be parsed as a CSS <color> value then it must be ignored, and the attribute must retain its previous value. [CSSCOLOR]

The shadowOffsetX and shadowOffsetY attributes specify the distance that the shadow will be offset in the positive horizontal and positive vertical distance respectively. Their values are in coordinate space units. They are not affected by the current transformation matrix.

When the context is created, the shadow offset attributes must initially have the value 0.

On getting, they must return their current value. On setting, the attribute being set must be set to the new value, except if the value is infinite or NaN, in which case the new value must be ignored.

The shadowBlur attribute specifies the level of the blurring effect. (The units do not map to coordinate space units, and are not affected by the current transformation matrix.)

When the context is created, the shadowBlur attribute must initially have the value 0.

On getting, the attribute must return its current value. On setting the attribute must be set to the new value, except if the value is negative, infinite or NaN, in which case the new value must be ignored.

Shadows are only drawn if the opacity component of the alpha component of the color of shadowColor is non-zero and either the shadowBlur is non-zero, or the shadowOffsetX is non-zero, or the shadowOffsetY is non-zero.

It is likely that this will change: browser vendors have indicated an interest in changing the processing model for shadows such that they only draw when the composition operator is "source-over" (the default). Read more...

When shadows are drawn, they must be rendered as follows:

1. Let A be an infinite transparent black bitmap on which the source image for which a shadow is being created has been rendered.

2. Let B be an infinite transparent black bitmap, with a coordinate space and an origin identical to A.

3. Copy the alpha channel of A to B, offset by shadowOffsetX in the positive x direction, and shadowOffsetY in the positive y direction.

4. If shadowBlur is greater than 0:

   1. Let σ be half the value of shadowBlur.

   2. Perform a 2D Gaussian Blur on B, using σ as the standard deviation.

   User agents may limit values of σ to an implementation-specific maximum value to avoid exceeding hardware limitations during the Gaussian blur operation.

5. Set the red, green, and blue components of every pixel in B to the red, green, and blue components (respectively) of the color of shadowColor.

6. Multiply the alpha component of every pixel in B by the alpha component of the color of shadowColor.

7. The shadow is in the bitmap B, and is rendered as part of the drawing model described below.

20 Drawing model

When a shape or image is painted, user agents must follow these steps, in the order given (or act as if they do):

1. Render the shape or image onto an infinite transparent black bitmap, creating image A, as described in the previous sections. For shapes, the current fill, stroke, and line styles must be honored, and the stroke must itself also be subjected to the current transformation matrix.

2. When shadows are drawn, render the shadow from image A, using the current shadow styles, creating image B.

3. When shadows are drawn, multiply the alpha component of every pixel in B by globalAlpha.

4. When shadows are drawn, composite B within the clipping region over the current scratch bitmap using the current composition operator.

5. Multiply the alpha component of every pixel in A by globalAlpha.

6. Composite A within the clipping region over the current scratch bitmap using the current composition operator.

When compositing onto the scratch bitmap, pixels that would fall outside of the scratch bitmap must be discarded.