最近在工作中需要解析点9图的头信息（chunk）的格式，读取拉伸坐标片段、padding信息，在网上找了一下没有相关信息，下面详解一下解析过程。
点9图科普
点9图的定义见官方文档：http://developer.android.com/guide/topics/graphics/2d-graphics.html#nine-patch
简单来说就是左边的点代表垂直拉伸的区域，上边的点代表水平拉伸的区域；右边的点代表文字等的垂直可可显示区域，下边的点代表文字等的水平可显示区域。所以，左上重合的区域就是拉伸区域；右下重合的区域就是显示区域。
Android的编译工具aapt会把标记了拉伸区域的点9图处理成一个普通的png，并把相关的信息放入png头部的meta-data区域，即chunk。这个过程详见frameworks/base/tools/aapt/Images.cpp的static status_t do_9patch(const char* imageName, image_info* image)
PNG文件结构
官方文档参见右边的链接：The Metadata in PNG files，摘抄部分如下
A PNG always starts with an 8-byte signature: 137 80 78 71 13 10 26 10 (decimal values). The remainder of the file consists a series of chunks beginning with an IHDR chunk and ending with an IEND chunk.
好了，了解了PNG的文件结构我们就可以取出chunk了，只需要按照文档的说明，按顺序取出Chunk data对应的信息即可。特别需要注意的是，读取字节流的时候需要采取大端方式。下面上代码
/**
* PNG Chunk struct
* <a href="http://dev.exiv2.org/projects/exiv2/wiki/The_Metadata_in_PNG_files">The Metadata in PNG files</a>
*
* +--------+---------+
* | Length | 4 bytes |
* +--------+---------+
* | Chunk | 4 bytes |
* | type | |
* +--------+---------+
* | Chunk | Length |
* | data | bytes |
* +--------+---------+
* | CRC | 4 bytes |
* +--------+---------+
*
* @param pngName
* @return chunk
* @throws IOException
*/
private byte[] loadNinePatchChunk(InputStream is) throws IOException {
IntReader reader = new IntReader(is, true);
// check PNG signature
// A PNG always starts with an 8-byte signature: 137 80 78 71 13 10 26 10 (decimal values).
if (reader.readInt() != 0x89504e47 || reader.readInt() != 0x0D0A1A0A) {
return null;
}
while (true) {
int length = reader.readInt();
int type = reader.readInt();
// check for nine patch chunk type (npTc)
if (type != 0x6E705463) {
reader.skip(length + 4/*crc*/);
continue;
}
return reader.readByteArray(length);
}
}
解析Chunk
点9图的Chunk是按照struct Res_png_9patch的结构组织的，详见frameworks/base/include/androidfw/ResourceTypes.h。
struct Res_png_9patch
{
Res_png_9patch() : wasDeserialized(false), xDivs(NULL),
yDivs(NULL), colors(NULL) { }
int8_t wasDeserialized;
int8_t numXDivs;
int8_t numYDivs;
int8_t numColors;
// These tell where the next section of a patch starts.
// For example, the first patch includes the pixels from
// 0 to xDivs[0]-1 and the second patch includes the pixels
// from xDivs[0] to xDivs[1]-1.
// Note: allocation/free of these pointers is left to the caller.
int32_t* xDivs;
int32_t* yDivs;
int32_t paddingLeft, paddingRight;
int32_t paddingTop, paddingBottom;
enum {
// The 9 patch segment is not a solid color.
NO_COLOR = 0x00000001,
// The 9 patch segment is completely transparent.
TRANSPARENT_COLOR = 0x00000000
};
// Note: allocation/free of this pointer is left to the caller.
uint32_t* colors;
// Convert data from device representation to PNG file representation.
void deviceToFile();
// Convert data from PNG file representation to device representation.
void fileToDevice();
// Serialize/Marshall the patch data into a newly malloc-ed block
void* serialize();
// Serialize/Marshall the patch data
void serialize(void* outData);
// Deserialize/Unmarshall the patch data
static Res_png_9patch* deserialize(const void* data);
// Compute the size of the serialized data structure
size_t serializedSize();
};
了解了struct Res_png_9patch，我们只需根据struct Res_png_9patch的结构解析上面步骤得到的chunk data即可。代码如下：
public static NinePatchChunk deserialize(byte[] data) {
ByteBuffer byteBuffer = ByteBuffer.wrap(data).order(ByteOrder.BIG_ENDIAN);
byte wasSerialized = byteBuffer.get();
// if (wasSerialized == 0) return null;
NinePatchChunk chunk = new NinePatchChunk();
chunk.mDivX = new int[byteBuffer.get()];
chunk.mDivY = new int[byteBuffer.get()];
chunk.mColor = new int[byteBuffer.get()];
checkDivCount(chunk.mDivX.length);
checkDivCount(chunk.mDivY.length);
// skip 8 bytes
byteBuffer.getInt();
byteBuffer.getInt();
chunk.mPaddings.left = byteBuffer.getInt();
chunk.mPaddings.right = byteBuffer.getInt();
chunk.mPaddings.top = byteBuffer.getInt();
chunk.mPaddings.bottom = byteBuffer.getInt();
// skip 4 bytes
byteBuffer.getInt();
readIntArray(chunk.mDivX, byteBuffer);
readIntArray(chunk.mDivY, byteBuffer);
readIntArray(chunk.mColor, byteBuffer);
return chunk;
}
NinePatchChunk.mDivX和NinePatchChunk.mDivY就是点9图的拉伸区域标识信息，NinePatchChunk.mPaddings是点9图的pading信息，这样点9图的Chunk信息就解析完毕了。
总结
全部代码见https://gist.github.com/luxiaoyu/085135ff88d7c57a18c5
调研这种SDK文档上没有详细信息的数据结构，需要从Android源码入手，耐心查看相关代码和注释，grep和find是好帮手。
————————————————
版权声明：本文为CSDN博主「鲁晓宇-Baidu」的原创文章，遵循CC 4.0 BY-SA版权协议，转载请附上原文出处链接及本声明。
原文链接：https://blog.csdn.net/u013365670/java/article/details/25415393