我们开发过程,基本需要自定义View,画一些自己的小插件出来
这需要我们掌握整个View的绘画过程和一些别的小技巧。
这里总结下整个View的源码中涉及到的一些绘制过程的核心部分,
之后再来看下整体的内容,毕竟整个源码有近2W1行,不是随便一时半会能搞定的,还是得下不少功夫。
起航 ------ 绘制流程
API:23
一般View的“生命周期”即绘画的流程像下面这样。
st=>start: View的绘画流程
op=>operation: measure()
op2=>operation: layout()
op3=>operation: draw()
e=>end: 结束
st->op->op2->op3->e
这个是一般的流程都这样,
- 我们的
measure
负责去测量View的Width
和Height
, - 然后我们的
layout
负责去确定其在父容器的位置, - 最后由
draw
来负责在屏幕上画内容。
但实际还有一些别的步骤流程,如这些函数由上一层来调用, 就像我们的Activity的onCreate
等!
不过现在先不提及。我们继续看各个阶段具体到底是做什么先。
measure
public final void measure(int widthMeasureSpec, int heightMeasureSpec) {
boolean optical = isLayoutModeOptical(this);
...
onMeasure(widthMeasureSpec, heightMeasureSpec);
...
}
protected void onMeasure(int widthMeasureSpec, int heightMeasureSpec) {
setMeasuredDimension(getDefaultSize(getSuggestedMinimumWidth(), widthMeasureSpec),
getDefaultSize(getSuggestedMinimumHeight(), heightMeasureSpec));
}
public static int getDefaultSize(int size, int measureSpec) {
int result = size;
int specMode = MeasureSpec.getMode(measureSpec);
int specSize = MeasureSpec.getSize(measureSpec);
switch (specMode) {
case MeasureSpec.UNSPECIFIED:
result = size;
break;
case MeasureSpec.AT_MOST:
case MeasureSpec.EXACTLY:
result = specSize;
break;
}
return result;
}
我们的measure函数是个final类型的,里面主要是调用了onMeasure函数,由他做具体测量。
这里需要补充一部分内容,关于MeasureSpec.EXACTLY
,MeasureSpec.AT_MOST
和MeasureSpec.UNSPECIFIED
-
EXACTLY:
这个词的意思是父容器已经检测出View的精确大小(eg:width=200dp/match_parent),这时我们的View的最终大小值就是specSize的值。 -
AT_MOST:
这个词的意思是父容器指定了一个大小(eg:width=wrap_content),这时我们的View的大小是要小于等于specSize的值,最终大小到底是多大,要看View的具体实现。 -
UNSPECIFIED:
这个词的意思是父容器不对View有任何大小的限制,需要多大就设置多大,但这一般是系统内部用来表示一种测量的状态。当然还有别的用处,例如我们的ScrollView
,他就可以用这个来告诉子View,大小无限,任意画。
上面的解释看起来这个View的MeasureSpec类型由我们的LayoutParams来设置,但实际这个MeasureSpec是由View和父容器一起决定的。这个好理解,例如我们的LinearLayout里面有个View,前者设置最高为200dp,后者为300dp,最终这个子View大小不由自己设置的300dp决定。具体的测量过程,下次再开贴说,就不插在这里了。我们继续主线
这样我们回看上面应该就好理解getDefaultSize()里面的到底是什么意思了。
在 MeasureSpec.UNSPECIFIED:
的状况下,大小是result = size;
,由传过来的参数觉得,我们看下具体做了什么
protected int getSuggestedMinimumHeight() {
return (mBackground == null) ? mMinHeight : max(mMinHeight, mBackground.getMinimumHeight());
}
protected int getSuggestedMinimumWidth() {
return (mBackground == null) ? mMinWidth : max(mMinWidth, mBackground.getMinimumWidth());
}
我们拿getSuggestedMinimumHeight()
来看下
里面含义就是:
-
如果我没背景,那么就是mMinHeight大小,这个值对应于我们写的
android:minHeight="20dp"
属性,他的默认值是0。case R.styleable.View_minWidth: mMinWidth = a.getDimensionPixelSize(attr, 0); break;
-
如果我有背景,那就选背景的最小高度和mMinHeight中最大的。
这个背景的getMinimumHeight()
内容是/** * Returns the minimum height suggested by this Drawable. If a View uses this * Drawable as a background, it is suggested that the View use at least this * value for its height. (There will be some scenarios where this will not be * possible.) This value should INCLUDE any padding. * * @return The minimum height suggested by this Drawable. If this Drawable * doesn't have a suggested minimum height, 0 is returned. */ public int getMinimumHeight() { final int intrinsicHeight = getIntrinsicHeight(); return intrinsicHeight > 0 ? intrinsicHeight : 0;
}
自带的解释已经很具体了,返回Drawable的最小高度,没有的话就返回0;可能有些奇怪,说得好像我们的Drawable可以没高是的。确实有些没有,例如我们在自定义一些我们的圆角的Button在不同点击效果时候用到的<shape>
标签写的背景,他就没有。
继续回主线
protected final void setMeasuredDimension(int measuredWidth, int measuredHeight) {
...
setMeasuredDimensionRaw(measuredWidth, measuredHeight);
}
private void setMeasuredDimensionRaw(int measuredWidth, int measuredHeight) {
mMeasuredWidth = measuredWidth;
mMeasuredHeight = measuredHeight;
mPrivateFlags |= PFLAG_MEASURED_DIMENSION_SET;
}
最后就设置了测量的大小,是的测量的大小,不是最终的大小,最终的大小还是需要根据实际做调整的。
这样我们的measure,测量过程就基本结束了。
一些题外话:
这里补充一个早年无知时候遇到的坑,那时候项目要求弄一个像下面这样的一个带有气泡框的进度条,
这里写图片描述那时候就直接类似于下面这样,继承View,然后重写onDraw函数,在里面绘制好整个样子。
public class BubbleProgressBar extends View {
public void onDraw(@NonNull Canvas canvas) {
//画进度和泡泡框
}
}
但画好后,遇到个问题,这个View居然自动填充满整个界面,我设置的是wrap_content
,感觉应该是系统帮我搞好,弄成很小的一个啊,怎么就那么大呢? 后来查了资料发现,如果我们是直接继承于View,那需要重写下那个measure函数,要不然他就会自动填满,为啥呢?回看那个getDefaultSize
函数
case MeasureSpec.AT_MOST:
case MeasureSpec.EXACTLY:
result = specSize;
break;
我们的wrap_content
就是那个AT_MOST
和EXACTLY
是同条路,实际就等于写了Match_parent
。
所以我们得根据情况来做判断,来给点指定大小
@Override
protected void onMeasure(int widthMeasureSpec, int heightMeasureSpec) {
int widthMode = MeasureSpec.getMode(widthMeasureSpec);
int heightMode = MeasureSpec.getMode(heightMeasureSpec);
int widthSize = MeasureSpec.getSize(widthMeasureSpec);
int heightSize = MeasureSpec.getSize(heightMeasureSpec);
if(heightMode==MeasureSpec.AT_MOST widthMode == MeasureSpec.AT_MOST ){
setMeasuredDimension(mOurDefalutHeight,mOurDefalutWidth);
}
...
}
现在想想,大概当年设计这个View类的人遇到了这个默认初始化大小应该是多大才合适的问题,所以干脆直接来个填充全局的方式。
前进 ------ Layout过程
看完了测量出界面的大小,我们需要开始下一步layout的过程了。
layout主要是用来确定View的位置的,具体如下
public void layout(int l, int t, int r, int b) {
...
int oldL = mLeft;
int oldT = mTop;
int oldB = mBottom;
int oldR = mRight;
boolean changed = isLayoutModeOptical(mParent) ?
setOpticalFrame(l, t, r, b) : setFrame(l, t, r, b);
if (changed || (mPrivateFlags & PFLAG_LAYOUT_REQUIRED) == PFLAG_LAYOUT_REQUIRED) {
onLayout(changed, l, t, r, b);
mPrivateFlags &= ~PFLAG_LAYOUT_REQUIRED;
ListenerInfo li = mListenerInfo;
if (li != null && li.mOnLayoutChangeListeners != null) {
ArrayList<OnLayoutChangeListener> listenersCopy =
(ArrayList<OnLayoutChangeListener>)li.mOnLayoutChangeListeners.clone();
int numListeners = listenersCopy.size();
for (int i = 0; i < numListeners; ++i) {
listenersCopy.get(i).onLayoutChange(this, l, t, r, b, oldL, oldT, oldR, oldB);
}
}
}
...
}
整个流程大致是先用setFrame()函数去设置我们的View的位置,然后调用onLayout来让服从其确定之元素的位置,由于这个onLayout做的是具体的布局工作,需要具体的继承的人去做,例如我们的LinearLayout有水平和垂直之分,所以在View中里面什么也没有。最后是调用监听函数,通知他们onLayoutChange()了。
protected boolean setFrame(int left, int top, int right, int bottom) {
boolean changed = false;
if (mLeft != left || mRight != right || mTop != top || mBottom != bottom) {
changed = true;
// Remember our drawn bit
int drawn = mPrivateFlags & PFLAG_DRAWN;
int oldWidth = mRight - mLeft;
int oldHeight = mBottom - mTop;
int newWidth = right - left;
int newHeight = bottom - top;
boolean sizeChanged = (newWidth != oldWidth) || (newHeight != oldHeight);
// Invalidate our old position
invalidate(sizeChanged);
mLeft = left;
mTop = top;
mRight = right;
mBottom = bottom;
mRenderNode.setLeftTopRightBottom(mLeft, mTop, mRight, mBottom);
if (sizeChanged) {
sizeChange(newWidth, newHeight, oldWidth, oldHeight);
}
...
}
return changed;
}
/**
* Called from layout when this view should
* assign a size and position to each of its children.
*
* Derived classes with children should override this method
* and call layout on each of their children.
*/
protected void onLayout(boolean changed, int left, int top, int right, int bottom) {
}
好了,基本的layout过程就这么结束了,我们的View的布局也就确定了。
接下来就看下draw过程了。
前进前进------画界面的Draw
这个画的过程,主要就是把View绘制到屏幕上去,根据写的注释,我们看到View的绘制过程有这里六个步骤。其中两个可以忽略的。
/*
* Draw traversal performs several drawing steps which must be executed
* in the appropriate order:
*
* 1. Draw the background
* 2. If necessary, save the canvas' layers to prepare for fading
* 3. Draw view's content
* 4. Draw children
* 5. If necessary, draw the fading edges and restore layers
* 6. Draw decorations (scrollbars for instance)
*/
继续的步骤如下:
public void draw(Canvas canvas) {
final int privateFlags = mPrivateFlags;
final boolean dirtyOpaque = (privateFlags & PFLAG_DIRTY_MASK) == PFLAG_DIRTY_OPAQUE &&
(mAttachInfo == null || !mAttachInfo.mIgnoreDirtyState);
mPrivateFlags = (privateFlags & ~PFLAG_DIRTY_MASK) | PFLAG_DRAWN;
// Step 1, draw the background, if needed
int saveCount;
if (!dirtyOpaque) {
drawBackground(canvas);
}
// skip step 2 & 5 if possible (common case)
final int viewFlags = mViewFlags;
boolean horizontalEdges = (viewFlags & FADING_EDGE_HORIZONTAL) != 0;
boolean verticalEdges = (viewFlags & FADING_EDGE_VERTICAL) != 0;
if (!verticalEdges && !horizontalEdges) {
// Step 3, draw the content
if (!dirtyOpaque) onDraw(canvas);
// Step 4, draw the children
dispatchDraw(canvas);
// Step 6, draw decorations (scrollbars)
onDrawScrollBars(canvas);
if (mOverlay != null && !mOverlay.isEmpty()) {
mOverlay.getOverlayView().dispatchDraw(canvas);
}
// we're done...
return;
}
/*
* Here we do the full fledged routine...
* (this is an uncommon case where speed matters less,
* this is why we repeat some of the tests that have been
* done above)
*/
... 画特效部分
}
我们再细看下各个步骤
private void drawBackground(Canvas canvas) {
final Drawable background = mBackground;
...
final int scrollX = mScrollX;
final int scrollY = mScrollY;
if ((scrollX | scrollY) == 0) {
background.draw(canvas);
} else {
canvas.translate(scrollX, scrollY);
background.draw(canvas);
canvas.translate(-scrollX, -scrollY);
}
}
然后这个onDraw和我们onLayout一样,需要自己写,里面空空如也
protected void onDraw(Canvas canvas) {
}
然后那个dispatchDraw()也是,这个需要我们自己做,但这个更多的是针对于ViewGroup类的包含子View的。这样Draw事件就传递给下面,遍历所有的子View元素的Draw方法,绘制完所有。
/**
* Called by draw to draw the child views. This may be overridden
* by derived classes to gain control just before its children are drawn
* (but after its own view has been drawn).
* @param canvas the canvas on which to draw the view
*/
protected void dispatchDraw(Canvas canvas) {
}
这样我们的Draw过程也就介绍了。
一些补充
看完一个完整的View的绘制过程,这里补充一些关于ViewGroup的内容
ViewGroup绘制过程中还需要让他的各个子View去绘制。
measureChildren()
protected void measureChildren(int widthMeasureSpec, int heightMeasureSpec) {
final int size = mChildrenCount;
final View[] children = mChildren;
for (int i = 0; i < size; ++i) {
final View child = children[i];
if ((child.mViewFlags & VISIBILITY_MASK) != GONE) {
measureChild(child, widthMeasureSpec, heightMeasureSpec);
}
}
}
这里看到,他对于那些除了设置为Gone不可见的,都进行了绘制。
不过有一个点引起我的兴趣,这个size
的大小不是取数组children
的大小,而是mChildrenCount
这个值。难道这背后有一个什么故事?查了下没什么结果。。。
protected void measureChild(View child, int parentWidthMeasureSpec,
int parentHeightMeasureSpec) {
final LayoutParams lp = child.getLayoutParams();
final int childWidthMeasureSpec = getChildMeasureSpec(parentWidthMeasureSpec,
mPaddingLeft + mPaddingRight, lp.width);
final int childHeightMeasureSpec = getChildMeasureSpec(parentHeightMeasureSpec,
mPaddingTop + mPaddingBottom, lp.height);
child.measure(childWidthMeasureSpec, childHeightMeasureSpec);
}
绘制的过程也是直接调用他们的measure函数去执行。在获取到子View的MeasureSpec时,具体是:
public static int getChildMeasureSpec(int spec, int padding, int childDimension) {
int specMode = MeasureSpec.getMode(spec);
int specSize = MeasureSpec.getSize(spec);
int size = Math.max(0, specSize - padding);
int resultSize = 0;
int resultMode = 0;
switch (specMode) {
// Parent has imposed an exact size on us
case MeasureSpec.EXACTLY:
if (childDimension >= 0) {
resultSize = childDimension;
resultMode = MeasureSpec.EXACTLY;
} else if (childDimension == LayoutParams.MATCH_PARENT) {
// Child wants to be our size. So be it.
resultSize = size;
resultMode = MeasureSpec.EXACTLY;
} else if (childDimension == LayoutParams.WRAP_CONTENT) {
// Child wants to determine its own size. It can't be
// bigger than us.
resultSize = size;
resultMode = MeasureSpec.AT_MOST;
}
break;
// Parent has imposed a maximum size on us
case MeasureSpec.AT_MOST:
if (childDimension >= 0) {
// Child wants a specific size... so be it
resultSize = childDimension;
resultMode = MeasureSpec.EXACTLY;
} else if (childDimension == LayoutParams.MATCH_PARENT) {
// Child wants to be our size, but our size is not fixed.
// Constrain child to not be bigger than us.
resultSize = size;
resultMode = MeasureSpec.AT_MOST;
} else if (childDimension == LayoutParams.WRAP_CONTENT) {
// Child wants to determine its own size. It can't be
// bigger than us.
resultSize = size;
resultMode = MeasureSpec.AT_MOST;
}
break;
// Parent asked to see how big we want to be
case MeasureSpec.UNSPECIFIED:
if (childDimension >= 0) {
// Child wants a specific size... let him have it
resultSize = childDimension;
resultMode = MeasureSpec.EXACTLY;
} else if (childDimension == LayoutParams.MATCH_PARENT) {
// Child wants to be our size... find out how big it should
// be
resultSize = 0;
resultMode = MeasureSpec.UNSPECIFIED;
} else if (childDimension == LayoutParams.WRAP_CONTENT) {
// Child wants to determine its own size.... find out how
// big it should be
resultSize = 0;
resultMode = MeasureSpec.UNSPECIFIED;
}
break;
}
return MeasureSpec.makeMeasureSpec(resultSize, resultMode);
}
这里面做的事情,主要的就是根据父容器的MeasureSpec同时结合View本身的LayoutParams来共同决定子View的MeasureSpec,所以子元素能用的大小就是父容器的尺寸减去padding
int specSize = MeasureSpec.getSize(spec);
int size = Math.max(0, specSize - padding);
前面在说View的时候也有提到过这个,具体的View的大小是需要和父容器协商的。
根据上面的内容的决定子View的大小的过程,我们可以总结出一个规律,就是如果我们设置了具体的大小(dp/px)那就是ChildSize,要不然是ParentSize除了UNPSECIFIED,
childParams \ parentParams | EXACTLY | AT_MOST | UNSPECIFIED |
---|---|---|---|
dp/px | EXACTLY - childSize | EXACTLY - childSize | EXACTLY - childSize |
match_parent | EXACTLY - parentSize | EXACTLY - parentSize | UNSPECIFIED - 0 |
wrap_content | EXACTLY - parentSize | EXACTLY - parentSize | UNSPECIFIED - 0 |
后记
一个View的绘制过程就这样结束了,也没太大负责的内容,但一个View里面的内容还是很多可以说的,
例如:
- 他内部的
post
机制,他可以让我们减少对Handler的使用。 - Touch事件的传递
- View的滑动
这些内容我们后面继续慢慢的补充吧。
另外这个View的调用者是ViewRoot
,他的具体实现是ViewRootImpl
,在他的performTraversal函数里面,执行了我们的view的整个绘制周期的调用
st=>start: performTraversals()
e=>end: 结束
op1=>operation: View.measure
op2=>operation: View.layout
op3=>operation: View.draw
cond1=>condition: 不用重新Measure?
cond2=>condition: 不用重新Layout?
cond3=>condition: 不用重新Draw?
st->cond1->cond2
cond1(yes)->cond2
cond1(no)->op1
cond1->cond2->cond3
cond2(yes)->cond3
cond2(no)->op2
cond3(yes)->e
cond3(no)->op3
更具体的调用流程如下:
performMeasure->measure:
measure->onMeasure:
onMeasure--> View.measure:
另外我们的layout和draw的套路类似,就不细写.
网友评论