需求

在这里插入图片描述

在这里插入图片描述

分析

  • 布局分为横竖屏
  • 涉及到视频窗口的大小、位置切换

通过观察需求原型图可得知,横竖屏切换可以简单分成7块区域

  • 4个视频窗口
  • 1个title,显示「XX号房间」
  • 1个ViewGroup,放置「头像,头像,头像,N个观众」
  • 另1个ViewGroup,放置聊天窗口相关

横竖屏切一开始我的思路是完全不使用系统的横竖屏切换,使用rotation来切换横竖屏,切换过程添加一个转换动画,如下图所示:

在这里插入图片描述

<p style="color: #999999; font-size: 12px; text-align: center;"> (请无视中间那个小眼睛) </p>

但因为横屏之后依然有聊天功能,不调用系统的旋转,输入法依然还是竖屏的形式弹出来的。暂时没有查到解决办法。无奈改为使用系统的横竖屏切换,切换时通知ViewGroup重新计算测量、布局子View。

实现

首先我们复写onMeasure方法,遍历子View测量,指定为我们计算后的宽高,Mode设置为EXACTLY。

override fun onMeasure(widthSpace: Int, heightSpace: Int) {
    val width = MeasureSpec.getSize(widthMeasureSpec)
    val height = MeasureSpec.getSize(heightMeasureSpec)

    isVertical = height > width

    if (!isVertical) {
        videosWidth = (width * 0.548f).toInt()
        horizontalSmallVideoWidth = (videosWidth / 3.0f).toInt()
        horizontalSmallVideoHeight = (horizontalSmallVideoWidth * 0.6803f).toInt()
        topicViewHeight = height - horizontalSmallVideoHeight - videoViewSpacing
    }

    for (i in 0 until childCount) {
        val child = getChildAt(i)

        val location = if (child.tag == null) {
            val location = createAndCalcCoordinates(child, i, width, height)
            if (location.fromLeft == 0 && location.fromRight == 0) location.run {
                fromLeft = toLeft
                fromTop = toTop
                fromRight = toRight
                fromBottom = toBottom
            }
            location
        } else {
            child.tag as ViewLocation
        }

        child.tag = location
        child.measure(
            MeasureSpec.makeMeasureSpec(location.right - location.left, MeasureSpec.EXACTLY),
            MeasureSpec.makeMeasureSpec(location.bottom - location.top, MeasureSpec.EXACTLY)
        )
    }

    setMeasuredDimension(
        MeasureSpec.makeMeasureSpec(width, MeasureSpec.EXACTLY),
        MeasureSpec.makeMeasureSpec(height, MeasureSpec.EXACTLY)
    )
}

其中ViewLocalion对象存储了View四个边的位置(left、top、right、bottom)。
首次加载时创建ViewLocation对象,并根据横竖屏计算每个View的坐标,最终对象会存储到View的tag中。再次触发onMeasure时不会重新计算布局,而是沿用之前的坐标数据。

计算子View位置的函数只会在横竖屏切换、用户点击切换视频位置或大小时调用。

onLayout中遍历子View,通知布局并传入对应的四边位置即可。

override fun onLayout(changed: Boolean, l: Int, t: Int, r: Int, b: Int) {
    for (i in 0 until childCount) {
        val child = getChildAt(i)
        val location = child.tag as ViewLocation

        child.layout(location.left, location.top, location.right, location.bottom)
    }
}

计算子View位置的算法可以点击这里查看,就不贴出来了。有一个需要注意的细节是:更新View坐标(除横竖屏切换)时,实际更改的是目标坐标,而非当前坐标,这样做的目的是为了实现动画效果。

当前坐标指的是在onMeasureonLayout中使用的变量:left、top、right、bottom。
目标坐标表示动画执行完毕后View的位置。

执行动画与坐标计算充分解耦,无论想新增怎样的动画(平移、缩放),只需要根据需求计算好左、上、右、下的位置,发送给动画执行的Runnable即可。

如切换大屏View的计算:

fun toEquallyDividedVideos() {
    if (isAnimRunning) {
        return
    }
    if (!isVertical/* || !isSmallMode*/) {
        return
    }
    isAnimRunning = true
    isSmallMode = false
    topicIndex = -1

    val videoWidth = measuredWidth.shr(1)
    val videoHeight = (videoWidth.toFloat() * 0.6882f).toInt()

    val arr = arrayOfNulls<ViewLocation>(childCount)
    for (i in 0 until childCount) {
        val location = getChildLocationAndResetFromLocations(i)
        arr[i] = location

        val remainder = (i % 2)
        when (i) {
            in 0..3 -> location.run {
                toLeft = remainder * videoWidth + remainder * videoViewSpacing.shr(1)
                toTop = (if (i >= 2) i / 2 else 0) * (videoViewSpacing + videoHeight) + titleHeight
                toRight = toLeft + videoWidth - ((i + 1) % 2) * videoViewSpacing.shr(1)
                toBottom = toTop + videoHeight
            }
            5 -> location.run {
                toTop = titleHeight + videoHeight.shl(1) + videoViewSpacing
                toBottom = toTop + titleHeight
            }
            6 -> location.run {
                toTop = titleHeight.shl(1) + videoHeight.shl(1) + videoViewSpacing
            }
        }
    }

    post(AnimRunnable(arr.map { it!! }.toTypedArray(), mDuration = 200L, isRotation = false))
}

最终效果

在这里插入图片描述